Learning the behavior of LSTM Networks for time series prediction
What it does
Create a LSTM network to predict bakery sales
How we built it
We built it with tensorflow keras and pycharm
Challenges we ran into
Missing datas, variations in the data format, LSTM categorizing, data stationarity
Accomplishments that we're proud of
We managed to prepare the data and were able to train the network.
What we learned
We learned that the preporcessing takes a lot of effort and we have split the dataset into time windows for better training results.
What's next for Time Series Prediction for Bakery Turnovers
Optimize the Dataset input variables (min-max, stationarity, etc.) and the LSTM Network. Furthermore we have to check, if other network types like VAR or transformer networks can improve our predictions.