Operation and Maintenance Recommendation Based on Probability Matrix Factorization with Tags for Relay Protection Equipment

The feature attribute set covering the device ontology, operation state, maintenance and other equipment life trajectory process is extracted, the relaying protection equipment tagging is established, and the personalized recommendation of equipment operation and maintenance is made, which can provide technical support for the operation and maintenance, maintenance and technical upgrading of the relaying protection equipment efficiently. This paper analyzes the basic attributes and behavior attributes of the relay protection equipment, and establishes the tagging system of the relay protection equipment. On this basis, a joint probability matrix factorization recommendation algorithm based on equipment tags nearest neighbor selection is proposed. By using the feature information of the equipment tag, the approximate feature vectors of the device and the item are established and combined into a joint probability matrix factorization. The algorithm based on neighborhood influence considers the relationship among devices, items and tags, weakens the sparsity of tag information and rating information, and improves the recommendation quality.