Research On Recommendation Method And Marketing Strategy Of Vertical Domain Integrating User Personality

With the increasingly rich functions of social media platforms,social media platforms have become the main competitive battlefield for enterprises to carry out network marketing and recommendation.The new generation is not only the main user group of social media,but also the main force of current consumption,so the new generation is the key to the future growth of the consumer market.Recommendation method is the core technology of network marketing.At present,the field of personalized recommendation tends to be saturated,and the recommended products and services are single and repeated,and the information of users is narrow,which makes it difficult to tap the potential demand of consumer groups.At the same time,there are a large number of user generated text on social platforms,which to some extent reflects the user’s personality traits and interest preferences,and interest can be mapped to the user’s attention to the vertical field,so the identification of user’s personality traits and vertical field preference analysis has become an extremely important aspect to mine the user’s consumption needs and tendencies.In order to solve the above problems,a vertical domain preference prediction and recommendation method integrating user personality traits for young groups is proposed,which mainly works from the following aspects:(1)Construct a multi-label personality trait recognition model.Since it is difficult to obtain user personality trait labels,data mining technology is firstly used to collect user-generated texts on Sina Weibo platform,and a personality trait recognition model is constructed based on Chinese pre-training model,neural network model and attention mechanism,so as to identify user personality traits based on usergenerated content and emotional characteristics.The overall validity of the model was verified by the validity of the classification model and the validity of the affective feature.(2)Construct the vertical domain recommendation model.A vertical domain recommendation method based on Deep FM model and user personality fusion was proposed by using the personality trait data identified by the model as user personality trait labels.The overall effectiveness of the recommendation method was verified by two experiments on the effectiveness of personality trait and recommendation model respectively.(3)Construct a map and propose marketing strategies.Through visualization software Gephi,based on the labeled and predicted vertical domain labels of user personality traits and tendencies,the vertical domain maps of users with different personality traits were constructed to show the co-occurrence relationship between different personality traits and vertical domains.Then,horizontal and vertical comparisons were made between the maps.On the basis of the vertical comparison,a vertical domain text topic orientation map was further constructed to dig out different personality traits with the same degree of preference for a vertical domain and the difference of demand orientation in this vertical domain based on user-generated texts.Finally,on the basis of the horizontal and vertical comparison of the map,the reasons for the differences are analyzed,and the corresponding marketing strategies are proposed.Based on the above work,the following conclusions can be drawn:(1)The proposed method of identifying user personality traits based on Chinese text indicates the feasibility and accuracy of identifying user personality traits based on user-generated text to a certain extent,and the integration of emotional factors is conducive to improving the effect of personality trait recognition.(2)The proposed vertical domain recommendation method based on usergenerated content and integrating personality traits shows that integrating user personality traits can improve the accuracy of vertical domain recommendation method to a certain extent.(3)Different personality traits have different preferences for different vertical fields,and there are different demand tendencies for the same vertical field.By analyzing the results of vertical domain recommendation experiment,this paper explores the relationship between different personality traits and vertical domain preference.Based on this,some suggestions are made for enterprises’ network marketing strategies,providing certain support for enterprises’ positioning of marketing objects and developing marketing strategies.On the one hand,the research results deepen the research of user personality trait recognition model under user-generated content and confirm the important role of emotional characteristics in personality trait recognition.On the other hand,psychology is introduced into the field of recommendation,which enriches the research perspective of network marketing strategy formulation based on recommendation algorithm.A vertical domain recommendation method is proposed,and a graph is constructed according to the experimental results for analysis,to explore the network marketing strategies for different users and enterprises in different vertical fields,and to provide management suggestions for their social media marketing activities.Improve the effect of network marketing.