Abstract: This paper studies the real-time recommendation of air conditioning operation based on behavior similarity, and proposes an optimized KNN algorithm Fo-KNN algorithm, which divides user features into different feature groups, and deeply studies the influence of each feature group on air conditioning operation, thus providing users with more accurate air conditioning operation recommended service. It can be seen from the experimental results that compared with the classical recommendation algorithm, the method has a great improvement in accuracy.

Key words: Big data, User behavior, Real time recommendation, Automatic control