Pretty new to machine learning, working on reducing the complexity of data. Basically the data has a current-voltage curve of a solar panel that is 250 pts of current and 250pts of voltage.
I am trying to reduce the data and convert into 7 or 10 feature which i will later use for classification into various faults.
I have used autoencoder to perform this but results are pretty bad? am i missing anything?
Is autoencoder a good technique for 1D array? [