2-m-logistic-regression.py

2-m-logistic-regression.py

import fileinput
import random
import logistic_regression

# Config
data_folder = "../data"
input_filename = "processed.iris.data"

# Filepaths
input_data_filepath = data_folder + "/" + input_filename

# Read in dataset
print "Inputting data from '" + input_data_filepath + "'"
rows = []
for row in fileinput.input(input_data_filepath):
    content = row.split(",")
    if len(content) > 1:
        content = map(float, content)
	rows.append(content)
print "Data read in."

# Split into X and Y, training and test
print "Generating test and training data."
train_proportion = 0.75
X = map(lambda x: x[:-1], rows)
Y = map(lambda x: x[-1], rows)
X_train = []
Y_train = []
X_test = []
Y_test = []
for n in range(len(X)):
   if random.random() < train_proportion:
       X_train.append(X[n])
       Y_train.append(Y[n])
   else:
       X_test.append(X[n])
       Y_test.append(Y[n])

# Establish optimisation paramaeters
feature_count = 4
lr = 0.001
iter = 20

# Run the model and generate predictions
print "Generate model and run training."
lr_model = logistic_regression.LogisticRegression(feature_count, lr, iter)
lr_model.fit(X_train, Y_train)
Y_hat = lr_model.predict(X_test)

# Compare and contrast:
print "Comparing Y_test to Y_hat."
for n in range(len(X_test)):
    print str(n) + " - " + str(Y_test[n]) + " - " + str(Y_hat[n])