A method includes obtaining feature data including one or more features and obtaining priority information corresponding to the one or more features of the feature data. The method includes, based on the priority information, dividing the one or more features into first features and second features, compressing the second features, and combining the first features with the compressed second features. The method includes obtaining preprocessed features by applying a combination of the first features and the compressed second features to a plurality of first machine learning models. The method includes obtaining, by applying the preprocessed features to a second machine learning model, probability values of a plurality of applications that are predicted to be executed on the electronic device. The method includes, based on the probability values, executing one or more applications from among the plurality of applications in the background process of the electronic device.