Quora Answer Classifier

solution.py

import sys
import numpy as np

f = sys.stdin
# f = open("../input00.txt")

[n, m] = map(int, f.readline().strip().split())

X = np.zeros((n, m), dtype="float")
Y = np.zeros(n)
train_ids = []

for i in xrange(n):
    line = f.readline().strip().split()
    train_ids.append(line[0])
    Y[i] = int(line[1])
    for feat in line[2:]:
        [feat_index, feat_val] = feat.split(":")
        X[i][int(feat_index) - 1] = float(feat_val)

q = int(f.readline().strip())

X_test = np.zeros((q, m), dtype="float")
test_ids = []

for i in xrange(q):
    line = f.readline().strip().split()
    test_ids.append(line[0])
    for feat in line[1:]:
        [feat_index, feat_val] = feat.split(":")
        X_test[i][int(feat_index) - 1] = float(feat_val)


# training

from sklearn import preprocessing
from sklearn import svm

xtrain_std = preprocessing.StandardScaler().fit_transform(X)
ytrain = Y
xtest_std = preprocessing.StandardScaler().fit_transform(X_test)

def learn():
    clf = svm.LinearSVC(C = 10**1)
    clf.fit(xtrain_std, ytrain)
    
    return clf
    
def test(clf):
    return clf.predict(xtest_std)


clf = learn()
ypred = test(clf)

for i in xrange(len(ypred)):
    print test_ids[i] + " " + ("+1" if (ypred[i] > 0) else "-1")