This repository is to demonstrate how timeseries analysis and prediction can be performed in the browser.
The current example test case uses weather data (temperature) and energy data (uv_index) from the open-meteo.com API, and compares current AI model predictions with open-meteo.com API forecasts.
[Regression predicts the next data point (very accurate), LSTM predicts the next data points (accurate)] https://codesolutions2.github.io/timeseries_analysis/index9.html
- Add a final custom model
- Smarter data stacking to forecast 24 data points for both LSTM and Regression, instead of one data point forecasting
- Try to improve processing time