I created an ensemble model using decision tree probability and Neural network probability to predict the type of theft that might occur in Chicago at a given time at a specific location.
I am getting an accuracy of 43.5% on the training data and 42.6% on the validation data. This is less than the probability of getting a head or tail on a coin flip.
What models can I try for this problem?
Is an accuracy of 43% acceptable in such kind of problems?
Hoping for a response.