Merge branch 'IT18174236(Content_Creation_Component)' into 'master'

It18174236(content creation component)

See merge request !7

Prediction Componant/Prediction/prediction.py

+import warnings;
+
+warnings.simplefilter('ignore')
+
+import pandas as pd
+import fbprophet
+from fbprophet import Prophet
+import json
+from prophet.serialize import model_to_json, model_from_json
+from matplotlib import pyplot
+
+df = pd.read_csv('dataset.csv')
+
+# data preprocessing
+df['Year'] = df['Time Date'].apply(lambda x: str(x)[-4:])
+df['Month'] = df['Time Date'].apply(lambda x: str(x)[-6:-4])
+df['Day'] = df['Time Date'].apply(lambda x: str(x)[:-6])
+df['ds'] = pd.DatetimeIndex(df['Year'] + '-' + df['Month'] + '-' + df['Day'])
+
+df = df.loc[(df['Code'] == 2667437) & (df['Topic'] == 'QLD_CW_ST0203')]
+df.drop(['Time Date', 'Code', 'Topic', 'Year', 'Month', 'Day'], axis=1, inplace=True)
+df.columns = ['y', 'ds']
+
+# sample print
+print(df.head())
+
+# train model
+m = Prophet(interval_width=0.95, daily_seasonality=True)
+model = m.fit(df)
+
+# save trained model
+with open('serialized_model.json', 'w') as fout:
+    json.dump(model_to_json(m), fout)  # Save model
+
+# future prediction (next 100 days)
+future = m.make_future_dataframe(periods=100, freq='D')
+forecast = m.predict(future)
+
+# print predicted data
+print(forecast)
+
+# show charts prediction
+model.plot(forecast)
+pyplot.show()
+plot1 = m.plot(forecast)