UNDERSTANDING FBPROPHET

Facebook Prophet: Time Series Forecasting

This README provides a step-by-step guide on using Facebook Prophet for time series forecasting. The code snippets provided demonstrate various aspects of the Prophet library, such as data loading, model instantiation, training, prediction, visualization, cross-validation, and performance evaluation.

Installation

Before using Prophet, ensure that you have Python installed on your system. You can install Prophet and its dependencies using the following commands:

pip install prophet

pip install pystan

Load the time series data

df = pd.read_csv('example_wp_log_peyton_manning.csv')

View the first 10 rows of the data

df.head(10)

View the last 10 rows of the data

df.tail(10)

Get the dimensions of the data (rows, columns)

df.shape

Obtain information about the data (data types, missing values, etc.)

df.info()

Generate descriptive statistics of the data

df.describe()

THE MODEL

Instantiate the model

model = Prophet()

Train the model

model.fit(df)

Generate future timestamps using the ds column

Use the make_future_dataframe method to generate future ds values

future = model.make_future_dataframe(periods=365)

View the last 5 rows of the future dataframe

future.tail(5)

Make predictions

future = model.predict(future)

View the first 5 rows of the predicted values

future.head(5)

View the last 3 rows of the predicted values with uncertainty intervals

future[['ds', 'yhat', 'yhat_lower', 'yhat_upper']].tail(3)

VISUALIZATION

Plot the predicted values

figure_1 = model.plot(future)

Plot the trend and components of the future dataframe (weekly and yearly)

figure_2 = model.plot_components(future)

CROSS VALIDATION

from prophet.diagnostics import cross_validation

Perform cross-validation

cv = cross_validation(model, initial='730 days', period='180 days', horizon='365 days')

View the first 5 rows of the cross-validation results

cv.head(5)

PERFORMANCE METRICS

from prophet.diagnostics import performance_metrics

Calculate performance metrics

pm = performance_metrics(cv)

View the first 5 rows of the performance metrics

pm.head(5)

VISUALIZING CROSS VALIDATION PERFORMANCE METRICS

from prophet.plot import plot_cross_validation_metric

Plot a specific performance metric (e.g., SMAPE)

figure_3 = plot_cross_validation_metric(cv, metric='smape')

HYPERPARAMETER TUNING

For detailed information on hyperparameter tuning in Prophet, refer to the official Prophet documentation. The documentation covers topics such as choosing the number of Fourier terms, adjusting trend flexibility, modifying seasonality prior scale, and diagnosing overfitting and underfitting.

By following the steps outlined in this README and referring to the official documentation, you can effectively utilize Facebook Prophet for time series forecasting and optimize your model's performance for your specific dataset.