Timeseries-Forecasting

Using TimeSeries Generator and LTSM

About

This is a Time series Forecasting LTSM model which trains on a series of data and then after analysing the dataset.

It tries to predict the upcoming trend of the dataset

Dataset

For training and testing this model, the data has been taken from theFederal Reserve Economic Data

Note

This model is trained and tested on two dataset:

• Retail Sales: Clothing and Clothing Accessory Stores

• Effective Federal Funds Rate

Graphical Representation

Retail Sales: Clothing and Clothing Accessory Stores

• Loss Vs Val_Loss

• Prediction on The Test set

Effective Federal Funds Rate

• Fed Rates Graph

• Loss Vs Val_Loss

• Predicted Trend

Update

Right Now, I am learning about the time series forcasting, as i will learn more the model will be tuned more

• Dataset stationarity is achieved

• Agumented Dickey Fuller Test is implemented to chech whether the dataset is stationary or not

• Various Methods like "Rolling mean", "Difference", "Seasonal Difference" is applied to the dataset to remove any seasonality or trend

Note

Inversion of dataset and log inverse dataset will be implmented as soon as I done with the code