nba_nn_keras.py

import pandas as pd
import numpy as np
from keras.models import Sequential
from keras.layers import Dense
from keras.wrappers.scikit_learn import KerasRegressor
from sklearn.model_selection import cross_val_score
from sklearn.model_selection import KFold
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
# from sklearn.pipeline import Pipeline
from sklearn.preprocessing import LabelEncoder

import keras

df = pd.read_excel("2015-2016_NBA.xlsx")


cols_categorical = ["MAIN REF", "CREW", "TEAMS", "VENUE"]

#encoding categorical
for col in cols_categorical:
	df[col] = df[col].astype("category")

df[cols_categorical] = df[cols_categorical].apply(lambda x: x.cat.codes)
# df[cols_categorical[0]].replace(to_replace={-1: 0}, inplace=True)

pred_col = ["OPENING TOTAL"]

data_cols = ["MAIN REF", "CREW", "TEAMS", "VENUE",
			 "3P"
			]

np.random.seed = 0

mask = round(len(df) * .8)
print("mask value {}".format(mask))
train = df.loc[:mask, :]
test = df.loc[mask:, :]
all_train = train.loc[:, data_cols]
all_test = train.loc[:, pred_col]
print("Training set length: {}".format(len(train)))
print("Test set length: {}".format(len(test)))

x_train = train.loc[:, data_cols]
y_train = train.loc[:, pred_col]

if len(y_train.columns) > 1:
	raise Exception("Something fucked up somewhere. Y has more than one pred_col")

x_test = test.loc[:, data_cols]
y_test = test.loc[:, pred_col]

def build_nn():
	base_model = Sequential()

	base_model.add(Dense(20, input_shape=(len(data_cols),), activation="relu")) #
	base_model.add(Dense(144, activation="relu"))
	base_model.add(Dense(64, activation="relu"))

	base_model.compile(loss='mean_squared_error', optimizer='adam',
			metrics=['accuracy'])
	print(base_model.summary())
	return base_model

TL_nba_model = KerasRegressor(build_fn=build_nn, epochs=50, verbose=False)
TL_nba_model.fit(x_train, y_train)

# kfold = KFold(n_splits=10, random_state=0)
# model_results = cross_val_score(TL_nba_model, all_train, all_test, cv=kfold)
# print("Lets see what we got. Basline accuracy %.2f (%.2f) MSE" % (results.mean(), results.std()))