In [4]:

    

from datetime import datetime
import numpy as np
import matplotlib.pyplot as plt
from time import time
import seaborn as sns
import pandas as pd
import scipy
import zipfile
from matplotlib.backends.backend_pdf import PdfPages


    

In [5]:

    

z = zipfile.ZipFile('train.csv.zip')


    

In [18]:

    

train = pd.read_csv(z.open('train.csv'), parse_dates='Dates')[['X', 'Y', 'Category']]


    

In [21]:

    

mark = np.random.rand(len(train)) < 0.8


    

In [23]:

    

knn_train = train[mark]
knn_test = train[~mark]


    

In [24]:

    

x = knn_train[['X', 'Y']]
y = knn_train['Category'].astype('category')


    

In [25]:

    

from sklearn.neighbors import KNeighborsClassifier


    

In [26]:

    

knn = KNeighborsClassifier(n_neighbors=10)


    

In [27]:

    

knn.fit(x, y)


    

    
Out[27]:





KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',
           metric_params=None, n_neighbors=10, p=2, weights='uniform')

In [28]:

    

xx = knn_test[['X', 'Y']]
yy = knn_test['Category'].astype('category')


    

In [46]:

    

zz = zipfile.ZipFile('test.csv.zip')
test = pd.read_csv(zz.open('test.csv'), parse_dates=['Dates'])


    

In [53]:

    

submit = pd.DataFrame({'Id': test.Id.tolist()})
p = knn.predict(test[['X', 'Y']])
for category in y.cat.categories:
    submit[category] = np.where(p == category, 1, 0)
    
submit.to_csv('k_nearest_neigbour.csv', index = False)