1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189
| from sqlalchemy import * from threading import Timer import signal, sys, time, threading import configparser from datetime import datetime
import pandas import pickle from sklearn import datasets, cross_validation, metrics from sklearn import preprocessing
from sklearn.linear_model import PassiveAggressiveRegressor from sklearn.svm import SVR
class PredictData: def __init__(self, partial_fit_flag=True): self.partial_fit_flag = partial_fit_flag self.regressor = None def fit(self, X_train, y_train): if self.partial_fit_flag: self.regressor.partial_fit(X_train, y_train) else: self.regressor.fit(X_train, y_train)
def train(self, X_train, y_train): if self.regressor == None: if self.partial_fit_flag: self.regressor = PassiveAggressiveRegressor(random_state=42) else: self.regressor = SVR(kernel='rbf', C=1000, gamma=0.1) self.fit(X_train, y_train) else: if self.partial_fit_flag: self.fit(X_train, y_train) def predict(self, X_test): return self.regressor.predict(X_test) def metrics_errror(self, y_predict, y_test): return metrics.mean_squared_error(y_predict,y_test)
def save_model(self, name): if self.regressor != None: with open(name, 'wb') as m: pickle.dump(self.regressor, m) def load_model(self, name): with open(name, 'rb') as m: self.regressor = pickle.load(m)
class ProcessData: def __init__(self, scaler_partialfit_flag=True): self.scaler_partialfit_flag = scaler_partialfit_flag self.scaler = None
def fit(self, X_train): self.scaler.partial_fit(X_train)
def train_test_split_scale(self, data, target, testSize=0): X_train, X_test, y_train, y_test = cross_validation.train_test_split(data, target, test_size=testSize, random_state=42) if self.scaler == None: self.scaler = preprocessing.MinMaxScaler(feature_range=(-1, 1)) self.fit(X_train) if len(X_test) == 0: X_train = self.scaler.transform(X_train) return X_train, y_train else: X_train = self.scaler.transform(X_train) X_test = self.scaler.transform(X_test) return X_train, y_train, X_test, y_test def validate_date(self,datetime_string): try: return datetime.strptime(datetime_string,"%Y-%m-%d %H:%M:%S") except ValueError: return False def set_dateTime(self,datetime_string): if self.validate_date(datetime_string) != False: configW = configparser.RawConfigParser() configW.optionxform = str configW.add_section('DB') configW.set('DB', 'LatestTime', datetime_string) with open('DB.ini', 'w') as configfile: configW.write(configfile) def get_dateTime(self): configR = configparser.RawConfigParser() configR.read("DB.ini") db_time=configR.get("DB","LatestTime") if self.validate_date(db_time) == False: return 0 else: return db_time def load_data_from_db(self, alwaysLatestFlag=False): pass def load_data_from_file(self, path): try: data = pandas.read_csv(path) X = data[['Date','Time','Global_reactive_power','Voltage','Global_intensity','Sub_metering_1','Sub_metering_2','Sub_metering_3']] y = data['Global_active_power'] D = pandas.to_datetime(X["Date"], format="%d/%m/%Y") X["Month"] = D.apply(lambda x: x.month) X = X.drop(["Date"], axis=1)
T = pandas.to_datetime(X["Time"], format="%H:%M:%S") X["Hour"] = T.apply(lambda x: x.hour) X = X.drop(["Time"], axis=1) return X,y except: print('ERROR load_data_from_file!!!' ) return [],[]
def save_scaler(self, scalerName): if self.scaler != None: with open(scalerName, 'wb') as s: pickle.dump(self.scaler, s)
def load_scaler(self, scalerName): with open(scalerName, 'rb') as s: self.scaler = pickle.load(s)
def Online_Learning_Train(): pd = ProcessData() X,y = pd.load_data_from_file('pw3.csv') Xt,Xyt = pd.train_test_split_scale(X,y) pdd = PredictData() pdd.train(Xt, Xyt) Xy_predict = pdd.predict(Xt) scoreX = pdd.metrics_errror(Xy_predict,Xyt) pd.save_scaler('scaler.pkl') pdd.save_model('model.pkl') print('MSE: {0:f}'.format(scoreX)) def Online_Learning_Test(): pd = ProcessData() pd.load_scaler('scaler.pkl') X,y = pd.load_data_from_file('pw5.csv') Xt,Xyt = pd.train_test_split_scale(X,y) pdd = PredictData() pdd.load_model('model.pkl') pdd.train(Xt, Xyt) Xy_predict = pdd.predict(Xt) scoreX = pdd.metrics_errror(Xy_predict,Xyt) print('MSE: {0:f}'.format(scoreX)) def Offline_Learning_Train(): pd = ProcessData() X,y = pd.load_data_from_file('pw.csv') Xt,Xyt = pd.train_test_split_scale(X,y) pdd = PredictData(False) pdd.train(Xt, Xyt) Xy_predict = pdd.predict(Xt) scoreX = pdd.metrics_errror(Xy_predict,Xyt) pd.save_scaler('scaler.pkl') pdd.save_model('model.pkl') print('MSE: {0:f}'.format(scoreX))
def Offline_Learning_Test(): pd = ProcessData() pd.load_scaler('scaler.pkl') X,y = pd.load_data_from_file('pw5.csv') Xt,Xyt = pd.train_test_split_scale(X,y) pdd = PredictData(False) pdd.load_model('model.pkl') Xy_predict = pdd.predict(Xt) scoreX = pdd.metrics_errror(Xy_predict,Xyt) print('MSE: {0:f}'.format(scoreX))
def Main(): Online_Learning_Train() Online_Learning_Test() ''' Offline_Learning_Train() Offline_Learning_Test() ''' if __name__ == '__main__': Main()
|