lesson5-nlp-Copy1.ipynb

lesson5-nlp-Copy1.ipynb

{
  "cells": [
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:13.183594Z",
          "start_time": "2018-02-18T08:21:08.861512Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "%reload_ext autoreload\n%autoreload 2\n%matplotlib inline\nfrom fastai import *\nfrom fastai.nlp import *\nimport matplotlib.pyplot as plt\nfrom IPython.display import HTML\nfrom sklearn.linear_model import LogisticRegression",
      "execution_count": 1,
      "outputs": []
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## IMDB dataset and the sentiment classification task"
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "The [large movie review dataset](http://ai.stanford.edu/~amaas/data/sentiment/) contains a collection of 50,000 reviews from IMDB. The dataset contains an even number of positive and negative reviews. The authors considered only highly polarized reviews. A negative review has a score ≤ 4 out of 10, and a positive review has a score ≥ 7 out of 10. Neutral reviews are not included in the dataset. The dataset is divided into training and test sets. The training set is the same 25,000 labeled reviews.\n\nThe **sentiment classification task** consists of predicting the polarity (positive or negative) of a given text.\n\nTo get the dataset, in your terminal run the following commands:\n\n`wget http://ai.stanford.edu/~amaas/data/sentiment/aclImdb_v1.tar.gz`\n\n`gunzip aclImdb_v1.tar.gz`\n\n`tar -xvf aclImdb_v1.tar`"
    },
    {
      "metadata": {
        "collapsed": true
      },
      "cell_type": "markdown",
      "source": "### Tokenizing and term document matrix creation"
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:13.208017Z",
          "start_time": "2018-02-18T08:21:13.185116Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "texts_labels_from_folders",
      "execution_count": 2,
      "outputs": [
        {
          "data": {
            "text/plain": "<function fastai.text.texts_labels_from_folders>"
          },
          "execution_count": 2,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:13.229901Z",
          "start_time": "2018-02-18T08:21:13.210071Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "PATH=\"data/aclImdb/\"\nnames = ['neg','pos']",
      "execution_count": 3,
      "outputs": []
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:13.257963Z",
          "start_time": "2018-02-18T08:21:13.231625Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "PATH",
      "execution_count": 4,
      "outputs": [
        {
          "data": {
            "text/plain": "'data/aclImdb/'"
          },
          "execution_count": 4,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:13.400505Z",
          "start_time": "2018-02-18T08:21:13.259403Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "%ls {PATH}",
      "execution_count": 5,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": "imdbEr.txt  imdb.vocab  README  \u001b[0m\u001b[01;34mtest\u001b[0m/  \u001b[01;34mtrain\u001b[0m/\r\n"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:13.550813Z",
          "start_time": "2018-02-18T08:21:13.403598Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "%ls {PATH}train",
      "execution_count": 6,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": "labeledBow.feat  \u001b[0m\u001b[01;34mpos\u001b[0m/    unsupBow.feat  urls_pos.txt\r\n\u001b[01;34mneg\u001b[0m/             \u001b[01;34munsup\u001b[0m/  urls_neg.txt   urls_unsup.txt\r\n"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:13.850754Z",
          "start_time": "2018-02-18T08:21:13.559444Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "%ls {PATH}train/pos | head",
      "execution_count": 7,
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": "0_9.txt\r\n10000_8.txt\r\n10001_10.txt\r\n10002_7.txt\r\n10003_8.txt\r\n10004_8.txt\r\n10005_7.txt\r\n10006_7.txt\r\n10007_7.txt\r\n10008_7.txt\r\nls: write error\r\n"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:13.876286Z",
          "start_time": "2018-02-18T08:21:13.853045Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "f'{PATH}train'",
      "execution_count": 8,
      "outputs": [
        {
          "data": {
            "text/plain": "'data/aclImdb/train'"
          },
          "execution_count": 8,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:22.374160Z",
          "start_time": "2018-02-18T08:21:13.878365Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "trn,trn_y = texts_labels_from_folders(f'{PATH}train',names)\nval,val_y = texts_labels_from_folders(f'{PATH}test',names)",
      "execution_count": 9,
      "outputs": []
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "Here is the text of the first review"
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:22.398740Z",
          "start_time": "2018-02-18T08:21:22.376699Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "trn_y.sum()",
      "execution_count": 10,
      "outputs": [
        {
          "data": {
            "text/plain": "12500"
          },
          "execution_count": 10,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:22.432250Z",
          "start_time": "2018-02-18T08:21:22.401109Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "val_y.sum()",
      "execution_count": 11,
      "outputs": [
        {
          "data": {
            "text/plain": "12500"
          },
          "execution_count": 11,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:22.464133Z",
          "start_time": "2018-02-18T08:21:22.434065Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "trn[0]",
      "execution_count": 12,
      "outputs": [
        {
          "data": {
            "text/plain": "'I had to compare two versions of Hamlet for my Shakespeare class and unfortunately I picked this version. Everything from the acting (the actors deliver most of their lines directly to the camera) to the camera shots (all medium or close up shots...no scenery shots and very little back ground in the shots) were absolutely terrible. I watched this over my spring break and it is very safe to say that I feel that I was gypped out of 114 minutes of my vacation. Not recommended by any stretch of the imagination.'"
          },
          "execution_count": 12,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:22.502438Z",
          "start_time": "2018-02-18T08:21:22.466683Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "trn_y[0]",
      "execution_count": 13,
      "outputs": [
        {
          "data": {
            "text/plain": "0"
          },
          "execution_count": 13,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "[`CountVectorizer`](http://scikit-learn.org/stable/modules/generated/sklearn.feature_extraction.text.CountVectorizer.html) converts a collection of text documents to a matrix of token counts (part of `sklearn.feature_extraction.text`)."
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:22.529518Z",
          "start_time": "2018-02-18T08:21:22.505802Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "veczr = CountVectorizer(tokenizer=tokenize)",
      "execution_count": 14,
      "outputs": []
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "`fit_transform(trn)` finds the vocabulary in the training set. It also transforms the training set into a term-document matrix. Since we have to apply the *same transformation* to your validation set, the second line uses just the method `transform(val)`. `trn_term_doc` and `val_term_doc` are sparse matrices. `trn_term_doc[i]` represents training document i and it contains a count of words for each document for each word in the vocabulary."
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:32.885335Z",
          "start_time": "2018-02-18T08:21:22.532545Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "trn_term_doc = veczr.fit_transform(trn)\nval_term_doc = veczr.transform(val)",
      "execution_count": 15,
      "outputs": []
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:32.904645Z",
          "start_time": "2018-02-18T08:21:32.887769Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "trn_term_doc",
      "execution_count": 16,
      "outputs": [
        {
          "data": {
            "text/plain": "<25000x75132 sparse matrix of type '<class 'numpy.int64'>'\n\twith 3749745 stored elements in Compressed Sparse Row format>"
          },
          "execution_count": 16,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:32.928926Z",
          "start_time": "2018-02-18T08:21:32.906392Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "trn_term_doc[0]",
      "execution_count": 17,
      "outputs": [
        {
          "data": {
            "text/plain": "<1x75132 sparse matrix of type '<class 'numpy.int64'>'\n\twith 69 stored elements in Compressed Sparse Row format>"
          },
          "execution_count": 17,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:33.017610Z",
          "start_time": "2018-02-18T08:21:32.932044Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "vocab = veczr.get_feature_names(); vocab[5000:5005]",
      "execution_count": 18,
      "outputs": [
        {
          "data": {
            "text/plain": "['aussie', 'aussies', 'austen', 'austeniana', 'austens']"
          },
          "execution_count": 18,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:33.042926Z",
          "start_time": "2018-02-18T08:21:33.020058Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "w0 = set([o.lower() for o in trn[0].split(' ')]); w0",
      "execution_count": 19,
      "outputs": [
        {
          "data": {
            "text/plain": "{'(all',\n '(the',\n '114',\n 'absolutely',\n 'acting',\n 'actors',\n 'and',\n 'any',\n 'back',\n 'break',\n 'by',\n 'camera',\n 'camera)',\n 'class',\n 'close',\n 'compare',\n 'deliver',\n 'directly',\n 'everything',\n 'feel',\n 'for',\n 'from',\n 'ground',\n 'gypped',\n 'had',\n 'hamlet',\n 'i',\n 'imagination.',\n 'in',\n 'is',\n 'it',\n 'lines',\n 'little',\n 'medium',\n 'minutes',\n 'most',\n 'my',\n 'not',\n 'of',\n 'or',\n 'out',\n 'over',\n 'picked',\n 'recommended',\n 'safe',\n 'say',\n 'scenery',\n 'shakespeare',\n 'shots',\n 'shots)',\n 'shots...no',\n 'spring',\n 'stretch',\n 'terrible.',\n 'that',\n 'the',\n 'their',\n 'this',\n 'to',\n 'two',\n 'unfortunately',\n 'up',\n 'vacation.',\n 'version.',\n 'versions',\n 'very',\n 'was',\n 'watched',\n 'were'}"
          },
          "execution_count": 19,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:33.065051Z",
          "start_time": "2018-02-18T08:21:33.044840Z"
        },
        "scrolled": true,
        "trusted": true
      },
      "cell_type": "code",
      "source": "len(w0)",
      "execution_count": 20,
      "outputs": [
        {
          "data": {
            "text/plain": "69"
          },
          "execution_count": 20,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:33.084266Z",
          "start_time": "2018-02-18T08:21:33.066561Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "veczr.vocabulary_['absurd']",
      "execution_count": 21,
      "outputs": [
        {
          "data": {
            "text/plain": "1297"
          },
          "execution_count": 21,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:33.107524Z",
          "start_time": "2018-02-18T08:21:33.086441Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "trn_term_doc[0,5000]",
      "execution_count": 22,
      "outputs": [
        {
          "data": {
            "text/plain": "0"
          },
          "execution_count": 22,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## Naive Bayes"
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "We define the **log-count ratio** $r$ for each word $f$:\n\n$r = \\log \\frac{\\text{ratio of feature $f$ in positive documents}}{\\text{ratio of feature $f$ in negative documents}}$\n\nwhere ratio of feature $f$ in positive documents is the number of times a positive document has a feature divided by the number of positive documents."
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:33.127098Z",
          "start_time": "2018-02-18T08:21:33.108908Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "def pr(y_i):\n    p = x[y==y_i].sum(0)\n    return (p+1) / ((y==y_i).sum()+1)",
      "execution_count": 23,
      "outputs": []
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:33.196920Z",
          "start_time": "2018-02-18T08:21:33.128415Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "x=trn_term_doc\ny=trn_y\n\nr = np.log(pr(1)/pr(0))\nb = np.log((y==1).mean() / (y==0).mean())",
      "execution_count": 24,
      "outputs": []
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "Here is the formula for Naive Bayes."
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:33.235179Z",
          "start_time": "2018-02-18T08:21:33.198793Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "pre_preds = val_term_doc @ r.T + b\npreds = pre_preds.T>0\n(preds==val_y).mean()",
      "execution_count": 25,
      "outputs": [
        {
          "data": {
            "text/plain": "0.81656"
          },
          "execution_count": 25,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "...and binarized Naive Bayes."
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:33.343511Z",
          "start_time": "2018-02-18T08:21:33.236645Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "x=trn_term_doc.sign()\nr = np.log(pr(1)/pr(0))\n\npre_preds = val_term_doc.sign() @ r.T + b\npreds = pre_preds.T>0\n(preds==val_y).mean()",
      "execution_count": 26,
      "outputs": [
        {
          "data": {
            "text/plain": "0.83016"
          },
          "execution_count": 26,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "### Logistic regression"
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "Here is how we can fit logistic regression where the features are the unigrams."
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:35.707315Z",
          "start_time": "2018-02-18T08:21:33.345293Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "m = LogisticRegression(C=1e8, dual=True)\nm.fit(x, y)\npreds = m.predict(val_term_doc)\n(preds==val_y).mean()",
      "execution_count": 27,
      "outputs": [
        {
          "data": {
            "text/plain": "0.83176"
          },
          "execution_count": 27,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:38.077055Z",
          "start_time": "2018-02-18T08:21:35.709283Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "m = LogisticRegression(C=1e8, dual=True)\nm.fit(trn_term_doc.sign(), y)\npreds = m.predict(val_term_doc.sign())\n(preds==val_y).mean()",
      "execution_count": 28,
      "outputs": [
        {
          "data": {
            "text/plain": "0.8548"
          },
          "execution_count": 28,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "...and the regularized version"
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:38.821670Z",
          "start_time": "2018-02-18T08:21:38.078757Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "m = LogisticRegression(C=0.1, dual=True)\nm.fit(x, y)\npreds = m.predict(val_term_doc)\n(preds==val_y).mean()",
      "execution_count": 29,
      "outputs": [
        {
          "data": {
            "text/plain": "0.84872"
          },
          "execution_count": 29,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:21:39.711370Z",
          "start_time": "2018-02-18T08:21:38.823314Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "m = LogisticRegression(C=0.1, dual=True)\nm.fit(trn_term_doc.sign(), y)\npreds = m.predict(val_term_doc.sign())\n(preds==val_y).mean()",
      "execution_count": 30,
      "outputs": [
        {
          "data": {
            "text/plain": "0.88404"
          },
          "execution_count": 30,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "### Trigram with NB features"
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "Our next model is a version of logistic regression with Naive Bayes features described [here](https://www.aclweb.org/anthology/P12-2018). For every document we compute binarized features as described above, but this time we use bigrams and trigrams too. Each feature is a log-count ratio. A logistic regression model is then trained to predict sentiment."
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:22:53.895640Z",
          "start_time": "2018-02-18T08:21:39.713400Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "veczr =  CountVectorizer(ngram_range=(1,3), tokenizer=tokenize, max_features=800000)\ntrn_term_doc = veczr.fit_transform(trn)\nval_term_doc = veczr.transform(val)",
      "execution_count": 31,
      "outputs": []
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:22:53.913147Z",
          "start_time": "2018-02-18T08:22:53.896853Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "trn_term_doc.shape",
      "execution_count": 32,
      "outputs": [
        {
          "data": {
            "text/plain": "(25000, 800000)"
          },
          "execution_count": 32,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:22:55.110825Z",
          "start_time": "2018-02-18T08:22:53.914846Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "vocab = veczr.get_feature_names()",
      "execution_count": 33,
      "outputs": []
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:22:55.153433Z",
          "start_time": "2018-02-18T08:22:55.114484Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "vocab[200000:200005]",
      "execution_count": 34,
      "outputs": [
        {
          "data": {
            "text/plain": "['by vast', 'by vengeance', 'by vengeance .', 'by vera', 'by vera miles']"
          },
          "execution_count": 34,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:22:55.323834Z",
          "start_time": "2018-02-18T08:22:55.155320Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "y=trn_y\nx=trn_term_doc.sign()\nval_x = val_term_doc.sign()",
      "execution_count": 35,
      "outputs": []
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:22:56.037758Z",
          "start_time": "2018-02-18T08:22:55.325889Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "r = np.log(pr(1) / pr(0))\nb = np.log((y==1).mean() / (y==0).mean())",
      "execution_count": 36,
      "outputs": []
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "Here we fit regularized logistic regression where the features are the trigrams."
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:23:01.388023Z",
          "start_time": "2018-02-18T08:22:56.039245Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "m = LogisticRegression(C=0.1, dual=True)\nm.fit(x, y);\n\npreds = m.predict(val_x)\n(preds.T==val_y).mean()",
      "execution_count": 37,
      "outputs": [
        {
          "data": {
            "text/plain": "0.905"
          },
          "execution_count": 37,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "Here is the $\\text{log-count ratio}$ `r`.  "
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "Here we fit regularized logistic regression where the features are the trigrams' log-count ratios."
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "### NB-Trigram "
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:23:01.415048Z",
          "start_time": "2018-02-18T08:23:01.390712Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "b",
      "execution_count": 38,
      "outputs": [
        {
          "data": {
            "text/plain": "0.0"
          },
          "execution_count": 38,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:23:01.611269Z",
          "start_time": "2018-02-18T08:23:01.417388Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "pre_preds = val_term_doc.sign() @ r.T + b\npreds = pre_preds.T>0\n(preds==val_y).mean()",
      "execution_count": 39,
      "outputs": [
        {
          "data": {
            "text/plain": "0.87912"
          },
          "execution_count": 39,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "### NBLogisitic Trigram"
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:23:04.764464Z",
          "start_time": "2018-02-18T08:23:01.615512Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "x_nb = x.multiply(r)\nm = LogisticRegression(dual=True, C=0.1)\nm.fit(x_nb, y);\n\nval_x_nb = val_x.multiply(r)\npreds = m.predict(val_x_nb)\n(preds.T==val_y).mean()",
      "execution_count": 40,
      "outputs": [
        {
          "data": {
            "text/plain": "0.91768"
          },
          "execution_count": 40,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:23:04.810027Z",
          "start_time": "2018-02-18T08:23:04.768461Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "def plot_coef_(m):\n    np.random.seed(42)\n    tmp = np.random.choice(m.coef_.flatten(), 10000)\n    plt.subplots(figsize=(12,9))\n    plt.subplot(211)\n    plt.title('Distribution of random sampling 1000 coefficients')\n    display(plt.scatter(np.arange(10000),tmp,alpha =0.03))\n    plt.subplot(212)\n    plt.hist(tmp,bins=100)\n    plt.show()\n\n  \n    ",
      "execution_count": 41,
      "outputs": []
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:23:05.454606Z",
          "start_time": "2018-02-18T08:23:04.812949Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "plot_coef_(m)",
      "execution_count": 42,
      "outputs": [
        {
          "data": {
            "text/plain": "<matplotlib.collections.PathCollection at 0x7fe498a09cf8>"
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": "<matplotlib.figure.Figure at 0x7fe49898fbe0>"
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## Some exploration on Tri-gram NBLogistic"
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:23:06.192534Z",
          "start_time": "2018-02-18T08:23:05.456187Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "x_nb = x.multiply(r)\nval_x_nb = val_x.multiply(r)",
      "execution_count": 43,
      "outputs": []
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:23:07.611134Z",
          "start_time": "2018-02-18T08:23:06.194238Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "m1 = LogisticRegression(dual=True, C=1e-5)\nm1.fit(x_nb, y);\n\npreds1 = m1.predict(val_x_nb)\npreds_prob1 = m1.predict_proba(val_x_nb)\n(preds1.T==val_y).mean()",
      "execution_count": 44,
      "outputs": [
        {
          "data": {
            "text/plain": "0.8454"
          },
          "execution_count": 44,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:23:10.061048Z",
          "start_time": "2018-02-18T08:23:07.614466Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "m2 = LogisticRegression(dual=True, C=0.1)\nm2.fit(x_nb, y);\n\npreds2 = m2.predict(val_x_nb)\npreds_prob2 = m2.predict_proba(val_x_nb)\n(preds2.T==val_y).mean()",
      "execution_count": 45,
      "outputs": [
        {
          "data": {
            "text/plain": "0.91768"
          },
          "execution_count": 45,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:23:20.054112Z",
          "start_time": "2018-02-18T08:23:10.063339Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "m3 = LogisticRegression(dual=True, C=1e10)\nm3.fit(x_nb, y);\n\npreds3 = m3.predict(val_x_nb)\npreds_prob3 = m3.predict_proba(val_x_nb)\n(preds3.T==val_y).mean()",
      "execution_count": 46,
      "outputs": [
        {
          "data": {
            "text/plain": "0.9144"
          },
          "execution_count": 46,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:23:20.229001Z",
          "start_time": "2018-02-18T08:23:20.057288Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "preds_prob = m1.predict_proba(val_x_nb)",
      "execution_count": 47,
      "outputs": []
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:23:20.264485Z",
          "start_time": "2018-02-18T08:23:20.235737Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "preds_prob1[:10]",
      "execution_count": 48,
      "outputs": [
        {
          "data": {
            "text/plain": "array([[0.51372, 0.48628],\n       [0.51603, 0.48397],\n       [0.50064, 0.49936],\n       [0.51959, 0.48041],\n       [0.50762, 0.49238],\n       [0.50714, 0.49286],\n       [0.51122, 0.48878],\n       [0.50424, 0.49576],\n       [0.51215, 0.48785],\n       [0.53089, 0.46911]])"
          },
          "execution_count": 48,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:23:20.300837Z",
          "start_time": "2018-02-18T08:23:20.268602Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "m1.intercept_,m2.intercept_,m3.intercept_",
      "execution_count": 49,
      "outputs": [
        {
          "data": {
            "text/plain": "(array([0.00172]), array([0.25098]), array([2.19659]))"
          },
          "execution_count": 49,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:23:21.111483Z",
          "start_time": "2018-02-18T08:23:20.302343Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "plot_coef_(m1)",
      "execution_count": 50,
      "outputs": [
        {
          "data": {
            "text/plain": "<matplotlib.collections.PathCollection at 0x7fe46c590e10>"
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": "<matplotlib.figure.Figure at 0x7fe481598a58>"
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:23:21.567511Z",
          "start_time": "2018-02-18T08:23:21.112944Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "plot_coef_(m2)",
      "execution_count": 51,
      "outputs": [
        {
          "data": {
            "text/plain": "<matplotlib.collections.PathCollection at 0x7fe47e351198>"
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": "<matplotlib.figure.Figure at 0x7fe4613d0048>"
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:23:22.023779Z",
          "start_time": "2018-02-18T08:23:21.569422Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "plot_coef_(m3)",
      "execution_count": 52,
      "outputs": [
        {
          "data": {
            "text/plain": "<matplotlib.collections.PathCollection at 0x7fe4927cd898>"
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": "<matplotlib.figure.Figure at 0x7fe479070748>"
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## fastai NBSVM++"
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:23:22.042947Z",
          "start_time": "2018-02-18T08:23:22.025097Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "sl=2000",
      "execution_count": 53,
      "outputs": []
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:23:24.311407Z",
          "start_time": "2018-02-18T08:23:22.044966Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "# Here is how we get a model from a bag of words\nmd = TextClassifierData.from_bow(trn_term_doc, trn_y, val_term_doc, val_y, sl)",
      "execution_count": 54,
      "outputs": []
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:23:24.498980Z",
          "start_time": "2018-02-18T08:23:24.475517Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "md.dotprod_nb_learner",
      "execution_count": 60,
      "outputs": [
        {
          "data": {
            "text/plain": "<bound method TextClassifierData.dotprod_nb_learner of <fastai.nlp.TextClassifierData object at 0x7fe496d05080>>"
          },
          "execution_count": 60,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## Exploration"
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T08:47:41.466918Z",
          "start_time": "2018-02-18T08:31:46.268353Z"
        },
        "scrolled": true,
        "trusted": true
      },
      "cell_type": "code",
      "source": "learner1 = md.dotprod_nb_learner(w_adj=0)\nlearner1.fit(0.02, 10, wds=1e-6, cycle_len=1)",
      "execution_count": 65,
      "outputs": [
        {
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "46212754b1b84beea98ffc0bbaf953e9",
              "version_major": 2,
              "version_minor": 0
            },
            "text/html": "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n<p>\n  If you're reading this message in the Jupyter Notebook or JupyterLab Notebook, it may mean\n  that the widgets JavaScript is still loading. If this message persists, it\n  likely means that the widgets JavaScript library is either not installed or\n  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n  Widgets Documentation</a> for setup instructions.\n</p>\n<p>\n  If you're reading this message in another frontend (for example, a static\n  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n  it may mean that your frontend doesn't currently support widgets.\n</p>\n",
            "text/plain": "HBox(children=(IntProgress(value=0, description='Epoch', max=10), HTML(value='')))"
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": "epoch      trn_loss   val_loss   <lambda>                     \n    0      0.098939   0.148703   0.919349  \n    1      0.046574   0.130798   0.923841                     \n    2      0.025853   0.123561   0.923937                     \n    3      0.019665   0.118941   0.924217                     \n    4      0.015709   0.116099   0.923737                     \n    5      0.013377   0.11357    0.924057                     \n    6      0.012092   0.112222   0.923378                     \n    7      0.012003   0.110748   0.923418                      \n    8      0.011729   0.109428   0.923018                      \n    9      0.011766   0.108934   0.921739                      \n\n"
        },
        {
          "data": {
            "text/plain": "[0.108934045, 0.9217391304347825]"
          },
          "execution_count": 65,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T09:43:41.845993Z",
          "start_time": "2018-02-18T09:28:44.531094Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "learner2 = md.dotprod_nb_learner(w_adj=0.4)\nlearner2.fit(0.02, 10, wds=1e-6, cycle_len=1)",
      "execution_count": 76,
      "outputs": [
        {
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "c4713707eede457aab1ccaa0400bc4ff",
              "version_major": 2,
              "version_minor": 0
            },
            "text/html": "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n<p>\n  If you're reading this message in the Jupyter Notebook or JupyterLab Notebook, it may mean\n  that the widgets JavaScript is still loading. If this message persists, it\n  likely means that the widgets JavaScript library is either not installed or\n  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n  Widgets Documentation</a> for setup instructions.\n</p>\n<p>\n  If you're reading this message in another frontend (for example, a static\n  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n  it may mean that your frontend doesn't currently support widgets.\n</p>\n",
            "text/plain": "HBox(children=(IntProgress(value=0, description='Epoch', max=10), HTML(value='')))"
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": "epoch      trn_loss   val_loss   <lambda>                     \n    0      0.024842   0.119798   0.915889  \n    1      0.01341    0.116518   0.919301                     \n    2      0.008977   0.11331    0.920285                      \n    3      0.007614   0.111032   0.920884                      \n    4      0.006354   0.109504   0.921124                      \n    5      0.006496   0.108125   0.921523                      \n    6      0.006082   0.107467   0.922123                      \n    7      0.005592   0.10662    0.921379                      \n    8      0.005814   0.106126   0.921579                      \n    9      0.005999   0.106116   0.92118                       \n\n"
        },
        {
          "data": {
            "text/plain": "[0.106115766, 0.9211796675191816]"
          },
          "execution_count": 76,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T09:58:20.442230Z",
          "start_time": "2018-02-18T09:43:41.847858Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "learner3 = md.dotprod_nb_learner(w_adj=0.5)\nlearner3.fit(0.02, 10, wds=1e-6, cycle_len=1)",
      "execution_count": 77,
      "outputs": [
        {
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "cd947840ccd94253b212ed2181e53643",
              "version_major": 2,
              "version_minor": 0
            },
            "text/html": "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n<p>\n  If you're reading this message in the Jupyter Notebook or JupyterLab Notebook, it may mean\n  that the widgets JavaScript is still loading. If this message persists, it\n  likely means that the widgets JavaScript library is either not installed or\n  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n  Widgets Documentation</a> for setup instructions.\n</p>\n<p>\n  If you're reading this message in another frontend (for example, a static\n  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n  it may mean that your frontend doesn't currently support widgets.\n</p>\n",
            "text/plain": "HBox(children=(IntProgress(value=0, description='Epoch', max=10), HTML(value='')))"
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": "epoch      trn_loss   val_loss   <lambda>                     \n    0      0.020293   0.114449   0.913491  \n    1      0.011694   0.111685   0.917287                     \n    2      0.008039   0.11023    0.918167                      \n    3      0.005585   0.109161   0.918366                      \n    4      0.005115   0.108145   0.918886                      \n    5      0.004686   0.107234   0.919126                      \n    6      0.005392   0.106574   0.919285                      \n    7      0.00522    0.106148   0.918806                      \n    8      0.005476   0.105371   0.919485                      \n    9      0.005262   0.10524    0.919925                      \n\n"
        },
        {
          "data": {
            "text/plain": "[0.105239674, 0.9199248721227622]"
          },
          "execution_count": 77,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T10:12:58.598470Z",
          "start_time": "2018-02-18T09:58:20.443678Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "learner4 = md.dotprod_nb_learner(w_adj=0.6)\nlearner4.fit(0.02, 10, wds=1e-6, cycle_len=1)",
      "execution_count": 78,
      "outputs": [
        {
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "f282f571b6dc453fb53140a3c4dd02c6",
              "version_major": 2,
              "version_minor": 0
            },
            "text/html": "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n<p>\n  If you're reading this message in the Jupyter Notebook or JupyterLab Notebook, it may mean\n  that the widgets JavaScript is still loading. If this message persists, it\n  likely means that the widgets JavaScript library is either not installed or\n  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n  Widgets Documentation</a> for setup instructions.\n</p>\n<p>\n  If you're reading this message in another frontend (for example, a static\n  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n  it may mean that your frontend doesn't currently support widgets.\n</p>\n",
            "text/plain": "HBox(children=(IntProgress(value=0, description='Epoch', max=10), HTML(value='')))"
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": "epoch      trn_loss   val_loss   <lambda>                     \n    0      0.020628   0.11301    0.908696  \n    1      0.01149    0.109541   0.91417                       \n    2      0.006226   0.108511   0.915689                      \n    3      0.005355   0.10802    0.915129                      \n    4      0.004894   0.107241   0.915409                      \n    5      0.004676   0.106758   0.91497                       \n    6      0.004485   0.10601    0.915889                      \n    7      0.004371   0.105789   0.915809                      \n    8      0.004077   0.105491   0.917128                      \n    9      0.004709   0.105029   0.916528                      \n\n"
        },
        {
          "data": {
            "text/plain": "[0.10502877, 0.9165281329923274]"
          },
          "execution_count": 78,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T10:29:08.949309Z",
          "start_time": "2018-02-18T10:12:58.599875Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "learner5 = md.dotprod_nb_learner(w_adj=1)\nlearner5.fit(0.02, 10, wds=1e-6, cycle_len=1)",
      "execution_count": 79,
      "outputs": [
        {
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "4cacf55b694449d2bd5278b1b1940710",
              "version_major": 2,
              "version_minor": 0
            },
            "text/html": "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n<p>\n  If you're reading this message in the Jupyter Notebook or JupyterLab Notebook, it may mean\n  that the widgets JavaScript is still loading. If this message persists, it\n  likely means that the widgets JavaScript library is either not installed or\n  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n  Widgets Documentation</a> for setup instructions.\n</p>\n<p>\n  If you're reading this message in another frontend (for example, a static\n  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n  it may mean that your frontend doesn't currently support widgets.\n</p>\n",
            "text/plain": "HBox(children=(IntProgress(value=0, description='Epoch', max=10), HTML(value='')))"
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": "epoch      trn_loss   val_loss   <lambda>                     \n    0      0.017895   0.110381   0.897706  \n    1      0.009778   0.106044   0.903341                      \n    2      0.005      0.105329   0.90438                       \n    3      0.005681   0.105277   0.905459                      \n    4      0.004463   0.104743   0.906298                      \n    5      0.004505   0.105114   0.905858                      \n    6      0.00476    0.10498    0.906418                      \n    7      0.003135   0.104847   0.906538                      \n    8      0.00261    0.104874   0.906738                      \n    9      0.004009   0.104657   0.906817                      \n\n"
        },
        {
          "data": {
            "text/plain": "[0.10465697, 0.9068174552429668]"
          },
          "execution_count": 79,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {
        "ExecuteTime": {
          "end_time": "2018-02-18T10:45:00.106133Z",
          "start_time": "2018-02-18T10:29:08.950776Z"
        },
        "trusted": true
      },
      "cell_type": "code",
      "source": "learner6 = md.dotprod_nb_learner(w_adj=0.1)\nlearner6.fit(0.02, 10, wds=1e-6, cycle_len=1)",
      "execution_count": 80,
      "outputs": [
        {
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "b36451ba17d74d42a4c5ebfe9dd366bf",
              "version_major": 2,
              "version_minor": 0
            },
            "text/html": "<p>Failed to display Jupyter Widget of type <code>HBox</code>.</p>\n<p>\n  If you're reading this message in the Jupyter Notebook or JupyterLab Notebook, it may mean\n  that the widgets JavaScript is still loading. If this message persists, it\n  likely means that the widgets JavaScript library is either not installed or\n  not enabled. See the <a href=\"https://ipywidgets.readthedocs.io/en/stable/user_install.html\">Jupyter\n  Widgets Documentation</a> for setup instructions.\n</p>\n<p>\n  If you're reading this message in another frontend (for example, a static\n  rendering on GitHub or <a href=\"https://nbviewer.jupyter.org/\">NBViewer</a>),\n  it may mean that your frontend doesn't currently support widgets.\n</p>\n",
            "text/plain": "HBox(children=(IntProgress(value=0, description='Epoch', max=10), HTML(value='')))"
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": "epoch      trn_loss   val_loss   <lambda>                     \n    0      0.061197   0.139025   0.921659  \n    1      0.030921   0.127095   0.923657                     \n    2      0.018416   0.120685   0.923737                     \n    3      0.012708   0.117009   0.923593                     \n    4      0.011846   0.114271   0.923673                      \n    5      0.009861   0.112293   0.923593                      \n    6      0.0096     0.110555   0.923402                      \n    7      0.010118   0.109478   0.923138                      \n    8      0.009676   0.108958   0.922794                      \n    9      0.009995   0.108002   0.922738                      \n\n"
        },
        {
          "data": {
            "text/plain": "[0.10800249, 0.9227381713554986]"
          },
          "execution_count": 80,
          "metadata": {},
          "output_type": "execute_result"
        }
      ]
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "## References"
    },
    {
      "metadata": {},
      "cell_type": "markdown",
      "source": "* Baselines and Bigrams: Simple, Good Sentiment and Topic Classification. Sida Wang and Christopher D. Manning [pdf](https://www.aclweb.org/anthology/P12-2018)"
    },
    {
      "metadata": {
        "trusted": true
      },
      "cell_type": "code",
      "source": "",
      "execution_count": null,
      "outputs": []
    }
  ],
  "metadata": {
    "gist": {
      "id": "",
      "data": {
        "description": "lesson5-nlp-Copy1.ipynb",
        "public": true
      }
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3",
      "language": "python"
    },
    "language_info": {
      "name": "python",
      "version": "3.6.4",
      "mimetype": "text/x-python",
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "pygments_lexer": "ipython3",
      "nbconvert_exporter": "python",
      "file_extension": ".py"
    },
    "toc": {
      "colors": {
        "hover_highlight": "#DAA520",
        "navigate_num": "#000000",
        "navigate_text": "#333333",
        "running_highlight": "#FF0000",
        "selected_highlight": "#FFD700",
        "sidebar_border": "#EEEEEE",
        "wrapper_background": "#FFFFFF"
      },
      "moveMenuLeft": true,
      "nav_menu": {
        "height": "104px",
        "width": "251px"
      },
      "navigate_menu": true,
      "number_sections": true,
      "sideBar": true,
      "threshold": 4,
      "toc_cell": false,
      "toc_section_display": "block",
      "toc_window_display": false,
      "widenNotebook": false
    },
    "varInspector": {
      "cols": {
        "lenName": 16,
        "lenType": 16,
        "lenVar": 40
      },
      "kernels_config": {
        "python": {
          "delete_cmd_postfix": "",
          "delete_cmd_prefix": "del ",
          "library": "var_list.py",
          "varRefreshCmd": "print(var_dic_list())"
        },
        "r": {
          "delete_cmd_postfix": ") ",
          "delete_cmd_prefix": "rm(",
          "library": "var_list.r",
          "varRefreshCmd": "cat(var_dic_list()) "
        }
      },
      "types_to_exclude": [
        "module",
        "function",
        "builtin_function_or_method",
        "instance",
        "_Feature"
      ],
      "window_display": false
    }
  },
  "nbformat": 4,
  "nbformat_minor": 2
}