- This is a Time Series Forecasting problem.
- The purpose of this project is to Forecast next timestamp given a sequence of history values.
- This module was built with Recurrent Neural Network (RNN) on top of Tensorflow and Keras.
Recurrent neural networks are a type of neural network that add the explicit handling of order in input observations.
This capability suggests that the promise of recurrent neural networks is to learn the temporal context of input sequences in order to make better predictions. That is, that the suite of lagged observations required to make a prediction no longer must be diagnosed and specified as in traditional time series forecasting, or even forecasting with classical neural networks. Instead, the temporal dependence can be learned, and perhaps changes to this dependence can also be learned.
-
Input: a sequence of history values
- 2017-01-01,339.7
- 2017-02-01,440.4
- 2017-03-01,315.9
- 2017-04-01,439.3
- 2017-05-01,401.3
- 2017-06-01,437.4
- 2017-07-01,575.5
- 2017-08-01,407.6
- 2017-09-01,682.0
- 2017-10-01,475.3
- 2017-11-01,581.3
- 2017-12-01,646.9
-
Output: the value on next timestamp
- 2018-01-01,678.5
- Example 1:
python3 train_predict.py ./data/sales.csv ./training_config.json
- Example 2:
python3 train_predict.py ./data/daily-minimum-temperatures-in-me.csv ./training_config.json