Score and recommendation method of medical resources based on latent semantic model

Abstract

The present invention provides a medical resource scoring and recommendation method based on a latent semantic model. The steps are as follows: the first step is to obtain the data of recommended resource items, and perform data filtering and cleaning; the second step is to use continuous action learning automata to train SVD++ Each parameter in the model; the third step, for the user, for all resource items in the set, calculate the predicted value of the user’s rating for each resource item based on the hidden semantic model; the fourth step, use the improved sorting algorithm to get TopN recommendation list. The present invention regards the predicted scoring results and the degree of correlation with the most relevant hidden class as the basis for generating the recommendation list, and also refers to the reference coefficient of the resource item itself, which can ensure that the recommended resource item is more likely to arouse the interest of the patient, and also It can be guaranteed that the resource item has a high enough score to be favored by the patient.

Description

technical field [0001] The present invention relates to the hidden semantic model score prediction and TopN recommendation application of medical resources in the field of medical big data, specifically, it relates to the medical treatment based on latent semantic model in medical big data based on the combination of learning automata and gradient descent algorithm Resource scoring and recommendation methods. Background technique [0002] Scoring and recommendation functions have been widely used in today's Internet products, such as Taobao's product recommendation, Facebook's friend recommendation, Baidu's news recommendation, etc. Product recommendations, personalized product recommendations, friend recommendations from various social networking sites, popular application recommendations, etc. Recommendations are truly ubiquitous and pervasive, and the proportion of value generated by recommendations has always been high. For the above recommendation problems, the recomme...

AI Technical Summary

Problems solved by technology

Overall, however, recommendation algorithms are still a subject for further research

[0003] After searching, no description or report of similar technology to the present invention has been found at present, and similar materials at home and abroad have not been collected yet.

Images