Search

CN-121996844-A - Cross-domain preference prediction recommendation method and system based on personality characteristics

CN121996844ACN 121996844 ACN121996844 ACN 121996844ACN-121996844-A

Abstract

The application discloses a cross-domain preference prediction recommendation method and a cross-domain preference prediction recommendation system based on personality characteristics, which are characterized in that personality original data and historical behavior data of a user are obtained, and the personality original data are converted into personality characteristic representations; based on personality characteristic representation, predicting preference scores of users on untouched items through a pre-constructed personality preference association model, generating a first recommendation list according to the preference scores, generating a second recommendation list according to historical behavior data, dynamically adjusting weights of the first recommendation list and the second recommendation list according to recommendation scenes and user feedback, carrying out deep fusion on the first recommendation list and the second recommendation list according to the weights, and generating a final recommendation list. According to the application, through the personality preference association model, accurate prediction and recommendation of people, information and articles which are never contacted by a user but are highly likely to be preferred are realized, and the problems of 'information cocoon houses' and 'cold start' generated by the traditional recommendation method only depending on the historical behaviors of the user are solved.

Inventors

  • QIAN YU
  • LI YICHENG
  • SUN ZHONGKAI

Assignees

  • 北京齿伦转动科技有限公司

Dates

Publication Date
20260508
Application Date
20260126

Claims (10)

  1. 1. The cross-domain preference prediction recommendation method based on personality characteristics is characterized by comprising the following steps of: Step 1, acquiring personality original data and historical behavior data of a user, and converting the personality original data into personality characteristic representation; step 2, predicting preference scores of users on non-contact items through a pre-constructed personality preference association model based on the personality characteristic representation; step 3, generating a first recommendation list according to the preference scores; step 4, generating a second recommendation list according to the historical behavior data; step 5, dynamically adjusting weights of the first recommendation list and the second recommendation list according to recommendation scenes and user feedback; And 6, carrying out depth fusion on the first recommendation list and the second recommendation list according to the weight, and generating a final recommendation list.
  2. 2. The method for predicting and recommending cross-domain preference based on personality characteristics according to claim 1, wherein in step 1, when the personality primary data of the user is obtained, the personality primary data is obtained from a psychological assessment questionnaire authorized by the user, language behavior analysis of the user, and multimodal interaction data.
  3. 3. The method for predicting and recommending cross-domain preference based on personality characteristics according to claim 1, wherein in step 1, the personality characteristics represent that a personality characteristic vector is adopted, and a psychological theory and a data modeling technology are adopted when the personality raw data is converted into the personality characteristic vector.
  4. 4. The method for predicting and recommending cross-domain preference based on personality characteristics according to claim 1, wherein in step 2, the personality preference association model is constructed by learning association relations between personality characteristics and cross-domain preferences in large-scale sample data through machine learning or statistical analysis algorithms.
  5. 5. The personality-based cross-domain preference prediction recommendation method of claim 1 wherein in step2 the preference scores include social preference scores, content preference scores, and merchandise preference scores.
  6. 6. The personality characteristic-based cross-domain preference prediction recommendation method according to claim 1, wherein in step 4, a collaborative filtering algorithm or a content recommendation algorithm is adopted when generating a second recommendation list according to the historical behavior data.
  7. 7. A cross-domain preference prediction recommendation system based on personality characteristics, comprising: The data acquisition module is used for acquiring personality original data and historical behavior data of a user and converting the personality original data into personality characteristic representation; the preference score calculation module is used for predicting the preference score of the user on the untouched item through a pre-constructed personality preference association model based on the personality characteristic representation; the first recommendation list generation module is used for generating a first recommendation list according to the preference scores; the second recommendation list generation module is used for generating a second recommendation list according to the historical behavior data; the weight adjustment module is used for dynamically adjusting the weights of the first recommendation list and the second recommendation list according to the recommendation scene and user feedback; And the fusion recommendation module is used for carrying out deep fusion on the first recommendation list and the second recommendation list according to the weight and generating a final recommendation list.
  8. 8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 6 when the computer program is executed.
  9. 9. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 6.
  10. 10. A computer program product comprising a computer program or instructions which, when executed by a processor, carries out the steps of the method according to any one of claims 1 to 6.

Description

Cross-domain preference prediction recommendation method and system based on personality characteristics Technical Field The application relates to the technical field of digital content distribution, in particular to a cross-domain preference prediction recommendation method and system based on personality characteristics. Background With the explosive growth of internet information, personalized recommendation systems have become core technologies in the fields of electronic commerce, digital content distribution, social networks and the like. Conventional recommendation systems rely primarily on historical behavioral data of users (e.g., browsing, clicking, purchasing, scoring records, etc.) for preference mining and prediction. However, the conventional recommended mode has two fundamental drawbacks associated with each other: Firstly, the information cocoon house effect is that the traditional recommendation system continuously recommends content which is highly similar to the past interests of the user, so that the range of the user contact information is continuously narrowed, and the user contact information is gradually trapped in the inherent interest preference. The method not only limits the possibility of users to explore new fields and expand cognitive boundaries, but also reduces the freshness and satisfaction of the users to the recommendation system in the long term. Secondly, the "cold start" problem is that for a new user, the conventional recommendation system cannot generate an effective personalized recommendation for it because it has not generated enough historical behavioral data. Also, for newly emerging content or merchandise on the platform that lacks user interaction data (i.e., "new item cold start"), the system is also difficult to reach accurately. In essence, the limitation of the prior art is that its recommendation logic is built entirely on top of the explicit behavior that the user "has occurred" and cannot effectively insights and quantify the underlying, steady psychological motivation that drives the user's behavior. Thus, conventional systems can only mine and iterate for the user's known interests, and simply fail to predict the user's potential preferences for areas or items that may exist from never-contacted, appearing frustrating in dealing with the expansion of interests and the context of new users/items. Therefore, the industry is in urgent need of a recommendation technology capable of breaking through the limitation of historical behavior data, understanding the intrinsic characteristics of users from a deeper level, and realizing cross-field interest prediction according to the intrinsic characteristics of users, so as to fundamentally relieve the problems of 'information cocoons' and 'cold start', and further improve the exploration capacity of a recommendation system and long-term experience of users. Disclosure of Invention Therefore, the application provides a cross-domain preference prediction recommendation method and system based on personality characteristics, which are used for solving the problems of information cocoon houses and cold start in the prior art. In order to achieve the above object, the present application provides the following technical solutions: in a first aspect, a cross-domain preference prediction recommendation method based on personality characteristics includes: Step 1, acquiring personality original data and historical behavior data of a user, and converting the personality original data into personality characteristic representation; step 2, predicting preference scores of users on non-contact items through a pre-constructed personality preference association model based on the personality characteristic representation; step 3, generating a first recommendation list according to the preference scores; step 4, generating a second recommendation list according to the historical behavior data; step 5, dynamically adjusting weights of the first recommendation list and the second recommendation list according to recommendation scenes and user feedback; And 6, carrying out depth fusion on the first recommendation list and the second recommendation list according to the weight, and generating a final recommendation list. Preferably, in step 1, when the personality data of the user is obtained, the personality data is obtained from a psychological assessment questionnaire authorized by the user, a language behavior analysis of the user, and multimodal interaction data. Preferably, in step 1, the personality characteristic representation adopts a personality characteristic vector, and psychological theory and data modeling technology are adopted when the personality original data is converted into the personality characteristic vector. Preferably, in step2, the personality preference association model is constructed by learning association relations between personality characteristics and cross-domain preferences in the large-scal