The Application of Latent Feature Model in Power Equipment Evaluation

Healthy running of power equipment is an important guarantee of stable operation of power system. To ensure the healthy operation of electrical equipment, periodic assessment of equipment status is required. Traditional neural network method has high evaluation accuracy and parallel processing capability. However the algorithm takes a long time to train and is difficult to converge. Latent Feature Model is an important evaluation method of e-commerce. It has the advantages of fast online calculation and high evaluation accuracy. The shortcoming of the model is that personal preferences have an impact on the assessment results. In this paper, smart meter is selected as the experimental object, and the Latent Feature Model is used to evaluate the health status of smart meter. In order to remove the influence of personal preference, this article adopts the k-means algorithm to cluster the influence factors which eliminate the assessment bias caused by personal preferences. Through the experiment of 1573 smart meters, the elapsed time of the Latent Feature Model is reduced by 29.34% compared with the traditional artificial neural network method, and the evaluation accuracy is only lowered by only 1.94%.