def prep_us_nfl_pred(input_df,
use_csv=False,
data_csv='data/output_dont_commit/reg_output.csv'):
if use_csv:
input_df = sf.load_model_data(data_csv)
# Model Run - K-Means Clustering - Data Preparation
# Adding Mean Squared Error Column
input_df['Squared Error'] = input_df.apply(lambda row: (
(row['Total points'] - row['Total points predicted'])**2),
axis=1)
# Dropping Unwanted Columns
us_columns_to_drop = ['Total points', 'Total points predicted']
us_input_df = input_df.drop(us_columns_to_drop, axis=1)
logger.info(sf.Color.BOLD + sf.Color.GREEN +
"Sample Clustering Input Data:" + sf.Color.END)
logger.info(us_input_df.head(3))
# Using Label Encoding to Rebase the Values in these Columns
us_input_df = MultiColumnLabelEncoder(
columns=['name', 'year', 'team']).fit_transform(us_input_df)
logger.info(sf.Color.BOLD + sf.Color.GREEN +
"Sample Data Post Label Encoding:" + sf.Color.END)
logger.info(us_input_df.head(3))
# Converting to NumPy Array
us_input_npa = us_input_df.as_matrix().astype(np.float)
return input_df, us_input_npa
def prep_us_nfl_pred(input_df, use_csv=False, data_csv='data/output_dont_commit/reg_output.csv'):
if use_csv:
input_df = sf.load_model_data(data_csv)
# Model Run - K-Means Clustering - Data Preparation
# Adding Mean Squared Error Column
input_df['Squared Error'] = input_df.apply(lambda row: ((row['Total points'] -
row['Total points predicted']) ** 2), axis=1)
# Dropping Unwanted Columns
us_columns_to_drop = ['Total points', 'Total points predicted']
us_input_df = input_df.drop(us_columns_to_drop, axis=1)
logger.info(sf.Color.BOLD + sf.Color.GREEN + "Sample Clustering Input Data:" + sf.Color.END)
logger.info(us_input_df.head(3))
# Using Label Encoding to Rebase the Values in these Columns
us_input_df = MultiColumnLabelEncoder(columns=['name', 'year', 'team']).fit_transform(us_input_df)
logger.info(sf.Color.BOLD + sf.Color.GREEN + "Sample Data Post Label Encoding:" + sf.Color.END)
logger.info(us_input_df.head(3))
# Converting to NumPy Array
us_input_npa = us_input_df.as_matrix().astype(np.float)
return input_df, us_input_npa
def prep_reg_nfl_pred(feature_scaling=False,
data_csv='data/nfl_pred_data.csv'):
nfl_df = sf.load_model_data(data_csv)
col_names = nfl_df.columns.tolist()
logger.info(sf.Color.BOLD + sf.Color.GREEN + "Column Names:" +
sf.Color.END)
logger.info(col_names)
to_show = col_names[:]
logger.info(sf.Color.BOLD + sf.Color.GREEN + "Sample Loaded Data:" +
sf.Color.END)
logger.info(nfl_df[to_show].head(3))
# Isolate Target Data
y = np.array(nfl_df['Total points'])
# Columns to Drop (For Features Data Frame)
to_drop = ['Total points']
nfl_feat_space = nfl_df.drop(to_drop, axis=1)
# Capturing Feature Names
feature_names = nfl_feat_space.columns.tolist()
logger.info(sf.Color.BOLD + sf.Color.GREEN + "Feature Names:" +
sf.Color.END)
logger.debug(feature_names)
# Using Label Encoding to Rebase the Values in these Columns
nfl_feat_space = MultiColumnLabelEncoder(
columns=['name', 'year', 'team']).fit_transform(nfl_feat_space)
logger.info(sf.Color.BOLD + sf.Color.GREEN +
"Sample Data Post Label Encoding:" + sf.Color.END)
logger.info(nfl_feat_space.head(3))
# Make NumPy Array
x = nfl_feat_space.as_matrix().astype(np.float)
# Handle Feature Scaling and Normalization
if feature_scaling:
scaler = StandardScaler()
x = scaler.fit_transform(x)
logger.info(sf.Color.BOLD + sf.Color.GREEN + "Sample Transformed Data:" +
sf.Color.END)
logger.info(x[0:3])
logger.info("Feature Space holds %d Observations and %d Features" %
x.shape)
return [x, y, nfl_df]
def prep_reg_nfl_pred(feature_scaling=False, data_csv='data/nfl_pred_data.csv'):
nfl_df = sf.load_model_data(data_csv)
col_names = nfl_df.columns.tolist()
logger.info(sf.Color.BOLD + sf.Color.GREEN + "Column Names:" + sf.Color.END)
logger.info(col_names)
to_show = col_names[:]
logger.info(sf.Color.BOLD + sf.Color.GREEN + "Sample Loaded Data:" + sf.Color.END)
logger.info(nfl_df[to_show].head(3))
# Isolate Target Data
y = np.array(nfl_df['Total points'])
# Columns to Drop (For Features Data Frame)
to_drop = ['Total points']
nfl_feat_space = nfl_df.drop(to_drop, axis=1)
# Capturing Feature Names
feature_names = nfl_feat_space.columns.tolist()
logger.info(sf.Color.BOLD + sf.Color.GREEN + "Feature Names:" + sf.Color.END)
logger.debug(feature_names)
# Using Label Encoding to Rebase the Values in these Columns
nfl_feat_space = MultiColumnLabelEncoder(columns=['name', 'year', 'team']).fit_transform(nfl_feat_space)
logger.info(sf.Color.BOLD + sf.Color.GREEN + "Sample Data Post Label Encoding:" + sf.Color.END)
logger.info(nfl_feat_space.head(3))
# Make NumPy Array
x = nfl_feat_space.as_matrix().astype(np.float)
# Handle Feature Scaling and Normalization
if feature_scaling:
scaler = StandardScaler()
x = scaler.fit_transform(x)
logger.info(sf.Color.BOLD + sf.Color.GREEN + "Sample Transformed Data:" + sf.Color.END)
logger.info(x[0:3])
logger.info("Feature Space holds %d Observations and %d Features" % x.shape)
return [x, y, nfl_df]
##################################################################################################################
if __name__ == "__main__":
start_time = time.time()
# Machine Learning Chosen Models
estimator = KMeans
# Model Run - Clustering - K-Means - Estimator Keywords = dict()
estimator_keywords = dict(init='k-means++', n_init=10, verbose=0)
# Model Run - K-Means Clustering - Data Preparation
# Load Input Data Frame
input_df = sf.load_model_data('data/output_dont_commit/reg_output.csv')
# Dropping Unwanted Columns
columns_to_drop = ['Total points', 'Total points predicted']
us_input_df = input_df.drop(columns_to_drop, axis=1)
# Converting to NumPy Array
us_input_npa = us_input_df.as_matrix().astype(np.float)
# Model Run - K-Means Clustering
us_kcluster_df = run_clustering(us_input_npa,
make_plots=False,
clf_class=KMeans,
min_cluster=3,
max_cluster=3,
**estimator_keywords)
##################################################################################################################
if __name__ == "__main__":
start_time = time.time()
# Machine Learning Chosen Models
estimator = KMeans
# Model Run - Clustering - K-Means - Estimator Keywords = dict()
estimator_keywords = dict(init='k-means++', n_init=10, verbose=0)
# Model Run - K-Means Clustering - Data Preparation
# Load Input Data Frame
input_df = sf.load_model_data('data/output_dont_commit/reg_output.csv')
# Dropping Unwanted Columns
columns_to_drop = ['Total points', 'Total points predicted']
us_input_df = input_df.drop(columns_to_drop, axis=1)
# Converting to NumPy Array
us_input_npa = us_input_df.as_matrix().astype(np.float)
# Model Run - K-Means Clustering
us_kcluster_df = run_clustering(us_input_npa, make_plots=False, clf_class=KMeans, min_cluster=3,
max_cluster=3, **estimator_keywords)
# Model Run - K-Means Clustering - Output Processing
# Combine Input & Output Data Frames
us_result_df = pd.concat([input_df, us_kcluster_df], axis=1)
