BTC_LSTM_predictionmodel.ipynb

gistfile1.txt

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting https://github.com/s4w3d0ff/python-poloniex/archive/v0.4.6.zip\n",
      "  Downloading https://github.com/s4w3d0ff/python-poloniex/archive/v0.4.6.zip\n",
      "\u001b[K     / 40kB 112kB/s\n",
      "Collecting requests (from poloniex==0.4.6)\n",
      "  Downloading requests-2.18.4-py2.py3-none-any.whl (88kB)\n",
      "\u001b[K    100% |################################| 92kB 2.1MB/s ta 0:00:01\n",
      "\u001b[?25hRequirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python2.7/dist-packages (from requests->poloniex==0.4.6)\n",
      "Collecting chardet<3.1.0,>=3.0.2 (from requests->poloniex==0.4.6)\n",
      "  Downloading chardet-3.0.4-py2.py3-none-any.whl (133kB)\n",
      "\u001b[K    100% |################################| 143kB 2.6MB/s ta 0:00:01\n",
      "\u001b[?25hCollecting idna<2.7,>=2.5 (from requests->poloniex==0.4.6)\n",
      "  Downloading idna-2.6-py2.py3-none-any.whl (56kB)\n",
      "\u001b[K    100% |################################| 61kB 3.7MB/s ta 0:00:011\n",
      "\u001b[?25hCollecting urllib3<1.23,>=1.21.1 (from requests->poloniex==0.4.6)\n",
      "  Downloading urllib3-1.22-py2.py3-none-any.whl (132kB)\n",
      "\u001b[K    100% |################################| 133kB 2.8MB/s ta 0:00:01\n",
      "\u001b[?25hInstalling collected packages: chardet, idna, urllib3, requests, poloniex\n",
      "  Running setup.py install for poloniex ... \u001b[?25ldone\n",
      "\u001b[?25hSuccessfully installed chardet-3.0.4 idna-2.6 poloniex-0.4.6 requests-2.18.4 urllib3-1.22\n"
     ]
    }
   ],
   "source": [
    "# install python-poloniex\n",
    "!pip install https://github.com/s4w3d0ff/python-poloniex/archive/v0.4.6.zip"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import poloniex"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import time\n",
    "\n",
    "# prep poloniex API\n",
    "polo = poloniex.Poloniex()\n",
    "\n",
    "# 5min(300 seconds)100days\n",
    "chart_data = polo.returnChartData('BTC_ETH', period=300, start=time.time()-polo.DAY*100, end=time.time())\n",
    "#poloniex source code which explain how interval data we can get\n",
    "#https://github.com/s4w3d0ff/python-poloniex"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import pandas\n",
    "import pandas as pd\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "\n",
    "price_data = pd.DataFrame([float(i.get('open')) for i in chart_data])\n",
    "mss = MinMaxScaler()\n",
    "input_dataframe = pd.DataFrame(mss.fit_transform(price_data))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "#devide into training data and validation data\n",
    "\n",
    "import numpy as np\n",
    "\n",
    "def _load_data(data, n_prev=50):\n",
    "    docX, docY = [], []\n",
    "    for i in range(len(data)-n_prev):\n",
    "        docX.append(data.iloc[i:i+n_prev].as_matrix())\n",
    "        docY.append(data.iloc[i+n_prev].as_matrix())\n",
    "    alsX = np.array(docX)\n",
    "    alsY = np.array(docY)\n",
    "    return alsX, alsY\n",
    "\n",
    "#trainning data is 70%. Validationdata is 30%\n",
    "def train_test_split(df, test_size=0.3, n_prev=50):\n",
    "    ntrn = round(len(df) * (1 - test_size))\n",
    "    ntrn = int(ntrn)\n",
    "    X_train, y_train = _load_data(df.iloc[0:ntrn], n_prev)\n",
    "    X_test, y_test = _load_data(df.iloc[ntrn:], n_prev)\n",
    "    return (X_train, y_train), (X_test, y_test)\n",
    "\n",
    "(X_train, y_train), (X_test, y_test) = train_test_split(input_dataframe)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting keras\n",
      "  Downloading Keras-2.1.3-py2.py3-none-any.whl (319kB)\n",
      "\u001b[K    100% |################################| 327kB 1.8MB/s ta 0:00:01\n",
      "\u001b[?25hRequirement already satisfied: numpy>=1.9.1 in /usr/local/lib/python2.7/dist-packages (from keras)\n",
      "Requirement already satisfied: scipy>=0.14 in /usr/local/lib/python2.7/dist-packages (from keras)\n",
      "Requirement already satisfied: six>=1.9.0 in /usr/local/lib/python2.7/dist-packages (from keras)\n",
      "Collecting pyyaml (from keras)\n",
      "  Downloading PyYAML-3.12.tar.gz (253kB)\n",
      "\u001b[K    100% |################################| 256kB 2.1MB/s ta 0:00:01\n",
      "\u001b[?25hBuilding wheels for collected packages: pyyaml\n",
      "  Running setup.py bdist_wheel for pyyaml ... \u001b[?25ldone\n",
      "\u001b[?25h  Stored in directory: /root/.cache/pip/wheels/2c/f7/79/13f3a12cd723892437c0cfbde1230ab4d82947ff7b3839a4fc\n",
      "Successfully built pyyaml\n",
      "Installing collected packages: pyyaml, keras\n",
      "Successfully installed keras-2.1.3 pyyaml-3.12\n"
     ]
    }
   ],
   "source": [
    "!pip install keras"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python2.7/dist-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  from ._conv import register_converters as _register_converters\n",
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers.core import Dense, Activation\n",
    "from keras.layers.recurrent import LSTM\n",
    "\n",
    "in_out_neurons = 1\n",
    "hidden_neurons = 300\n",
    "length_of_sequences = 50\n",
    "\n",
    "model = Sequential()\n",
    "model.add(LSTM(hidden_neurons, batch_input_shape=(None, length_of_sequences, in_out_neurons), return_sequences=False))\n",
    "model.add(Dense(in_out_neurons))\n",
    "model.add(Activation(\"linear\"))\n",
    "model.compile(loss=\"mean_squared_error\", optimizer=\"adam\",)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers.core import Dense, Activation\n",
    "from keras.layers.recurrent import LSTM\n",
    "\n",
    "in_out_neurons = 1\n",
    "hidden_neurons = 300\n",
    "length_of_sequences = 50\n",
    "\n",
    "model = Sequential()\n",
    "model.add(LSTM(hidden_neurons, batch_input_shape=(None, length_of_sequences, in_out_neurons), return_sequences=False))\n",
    "model.add(Dense(in_out_neurons))\n",
    "model.add(Activation(\"linear\"))\n",
    "model.compile(loss=\"mean_squared_error\", optimizer=\"adam\",)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 18099 samples, validate on 2011 samples\n",
      "Epoch 1/10\n",
      "18099/18099 [==============================] - 45s 3ms/step - loss: 0.0041 - val_loss: 0.0033\n",
      "Epoch 2/10\n",
      "18099/18099 [==============================] - 45s 2ms/step - loss: 2.5816e-04 - val_loss: 3.4416e-04\n",
      "Epoch 3/10\n",
      "18099/18099 [==============================] - 42s 2ms/step - loss: 2.2754e-05 - val_loss: 2.9926e-05\n",
      "Epoch 4/10\n",
      "18099/18099 [==============================] - 41s 2ms/step - loss: 1.2393e-05 - val_loss: 2.4886e-05\n",
      "Epoch 5/10\n",
      "18099/18099 [==============================] - 42s 2ms/step - loss: 1.1886e-05 - val_loss: 2.4197e-05\n",
      "Epoch 6/10\n",
      "18099/18099 [==============================] - 47s 3ms/step - loss: 1.1829e-05 - val_loss: 2.4798e-05\n"
     ]
    }
   ],
   "source": [
    "from keras.callbacks import EarlyStopping\n",
    "\n",
    "early_stopping = EarlyStopping(monitor='val_loss', mode='auto', patience=0)\n",
    "history = model.fit(X_train, y_train, batch_size=600, epochs=10, validation_split=0.1, callbacks=[early_stopping])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f822a1e4cd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "from matplotlib import pyplot as plt\n",
    "\n",
    "pred_data = model.predict(X_train)\n",
    "plt.plot(y_train, label='train')\n",
    "plt.plot(pred_data, label='pred')\n",
    "plt.legend(loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f822a17f3d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pred_data = model.predict(X_test)\n",
    "plt.plot(y_test, label='test')\n",
    "plt.plot(pred_data, label='pred')\n",
    "plt.legend(loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f822a0bd350>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pred_data = model.predict(X_test)\n",
    "plt.plot(y_test, label='test')\n",
    "plt.plot(pred_data, label='pred')\n",
    "plt.legend(loc='upper left')\n",
    "plt.xlim(7800, 8100)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}