Manufacturing Qaulity Prediction with TimeSeries data and discrete quality measurements

12 Views Asked by At

I have timeseries data of a manufacturing process for 18 batches, each batch have 2000 rows of timeseries data for 6 variables. I also have 5 input quality measurements taken once per every batch and trying to predict a final product quality that is measured once per batch. Below is a sample format of all the data I have. Can I know what kind of models can I build with this data and how do I treat timeseries data to come up with a predictive model that combines the information from both time series data and input conditions. I have already tried some models converting the timeseries data to wide format making every timestamp a variable and merged the input conditions to make a PLS predictive model, that has 50% accuracy. I would like to explore machine learning methods to model this data but I am new to machine learning. Any ideas are appreciated. Thanks enter image description here

0

There are 0 best solutions below