# In ensemble learning, does accuracy increase depending on the number of models you want to combine?

I want to predict using the same model as multivariate time series data in a time series prediction problem.

Example:

pa = model predict result(a)
pb = model predict result(b)
pc = model predict result(c)
...
model ensemble([pa, pb, pc,...]) -> predict(y)


Can I expect a better performance of our model by using a model ensemble with more kinds of time series data here?

• Do you mean that would be a single model, but you have multiple timeseries with same target and you plan to ensemble prediction for these timeseries? Apr 1 '21 at 14:15
• Yes. I want to make predictions with multiple time series data of different kinds with the same model.
– KYH
Apr 1 '21 at 14:45

Yes, you can. Let's say you have 5 classes named a,b,c,d,e. You fit your data into a SVM Classifier and a Random Forest Classifier. Assume that, SVM classified "a" and "b" class well and RFC classified "c","d","e" well. So, ensembling these two models is going to increase accuracy dramatically. Ensemble learning is really good when a model generating many "False-Positive" and "False-Negative". You can also use weighted ensemble learning methods. You can set greater weights for reliable models & lower weights for untrustable models.