Create Niepotwierdzony 497997.crdownload

Niepotwierdzony 497997.crdownload

@@ -0,0 +1,121 @@
+# -*- coding: utf-8 -*-
+"""prophet.ipynb
+
+Automatically generated by Colaboratory.
+
+Original file is located at
+ https://colab.research.google.com/drive/1TjOXxO9UYYoDEdEIjyJqlDqO7PvZGeE9
+"""
+
+import pandas as pd
+import numpy as np
+import math
+
+!pip install sklearn-ts==0.0.5
+
+"""Load data"""
+
+covid = pd.read_csv("https://raw.githubusercontent.com/owid/covid-19-data/master/public/data/owid-covid-data.csv")
+#covid.head(2)
+
+target = 'new_cases'
+h = 14
+
+dataset = covid[(covid['location']=='World')].copy()[[target, 'date']]
+dataset[[target]].plot()
+
+# prepare features
+features = ['year', 'month', f'{h}_lag', f'{h}_lag_rolling', 'dayofweek', 'intercept', 'trend', 'log']
+categorical_features = ['year', 'month', 'dayofweek']
+numerical_features = [f'{h}_lag', f'{h}_lag_rolling', 'intercept', 'trend', 'log']
+
+dataset['date'] = pd.to_datetime(dataset['date'])
+dataset.index = dataset['date']
+dataset['month'] = dataset['date'].dt.month
+dataset['year'] = dataset['date'].dt.year
+dataset['dayofweek'] = dataset['date'].dt.dayofweek
+dataset[f'{h}_lag'] = dataset[target].shift(h)
+dataset[f'rolling_{target}'] = dataset[target].rolling(window=h).mean()
+dataset[f'{h}_lag_rolling'] = dataset[f'rolling_{target}'].shift(h)
+dataset['intercept'] = 1
+dataset['trend'] = range(dataset.shape[0])
+dataset['log'] = dataset['trend'].apply(lambda x: math.log(x+1))
+dataset = dataset[['date', target] + numerical_features +categorical_features]
+dataset = dataset.dropna()
+
+from sklearn_ts.validator import check_model
+
+! pip install prophet
+from prophet import Prophet
+
+df = dataset.copy()
+df['ds'] = df['date']
+df['y'] = df[target]
+m = Prophet(growth='linear', yearly_seasonality=True, weekly_seasonality=True, daily_seasonality=True)
+m.add_regressor(f'{h}_lag_rolling')
+m.fit(df)
+
+#future = m.make_future_dataframe(periods=14)
+forecast = m.predict(df)
+forecast[['ds', 'yhat', 'yhat_lower', 'yhat_upper']].tail()
+
+fig1 = m.plot(forecast)
+
+fig = m.plot_components(forecast)
+
+from prophet.utilities import regressor_coefficients
+
+regressor_coefficients(m)
+
+from sklearn.base import BaseEstimator, RegressorMixin
+from prophet.utilities import regressor_coefficients
+
+
+class ProphetRegressor(BaseEstimator, RegressorMixin):
+ # https://facebook.github.io/prophet/docs/quick_start.html#python-api
+
+ def __init__(self, target='new_cases', features=['date', f'{h}_lag_rolling'], regressors=[f'{h}_lag_rolling']):
+ self.target = target
+ self.features = features
+ self.regressors = regressors
+
+ self.model = None
+ self.predictions= None
+
+ def fit(self, X, y):
+ df = pd.DataFrame(X, columns=self.features)
+
+ df['ds'] = df['date']
+ df['y'] = y.values
+
+ m = Prophet(growth='linear', yearly_seasonality=True, weekly_seasonality=True, daily_seasonality=True)
+ for regressor in self.regressors:
+ m.add_regressor(regressor)
+ m.fit(df)
+
+ self.model = m
+ self.feature_importances_ = [None] + regressor_coefficients(m)['coef'].tolist() # place for date
+ return self
+
+ def predict(self, X):
+ df = pd.DataFrame(X, columns=self.features)
+ df['ds'] = df['date']
+ predictions = self.model.predict(df)
+ self.predictions = predictions[['ds', 'yhat_lower', 'yhat_upper']].rename(columns={'yhat_lower': 'pi_lower', 'yhat_upper': 'pi_upper'})
+ return predictions['yhat'].values
+
+ def get_params(self, deep=True):
+ return {"target": self.target, 'regressors': self.regressors, 'features': self.features}
+
+ def set_params(self, **parameters):
+ for parameter, value in parameters.items():
+ setattr(self, parameter, value)
+ return self
+
+params = {'regressors': [[f'{h}_lag_rolling']]}
+regressor = ProphetRegressor()
+
+results = check_model(regressor, params, dataset, features=['date', f'{h}_lag_rolling'])
+results['best_params']
+
+results['model'].named_steps['regressor'].predictions