Personalized Book Recommendation Model Research Merged With Individual Personality Trait

According to the rapid development of Internet, more and more e-book reading web sites arise, the amount of books’and users’data grows faster and faster, which makes the influence of data’s sparseness and cold start problems in book recommendation system becomes more crucial, as a result, recommendation system’s quality rapidly deteriorates. Due to the increasing amount of data, book scoring matrix becomes more sparse; As time goes on, the users’reading preferences change simultaneously, therefore even on the same book, it will have different evaluation based on different individual background factors, namely, individual background factors drift, so the locating data set is not precise enough; in addition, personality trait is another major factor affecting the users’behaviors. So how to improve the sparseness of book scoring matrix, and how to locate effective target data set in the process of recommendation is the focus of improving book recommendation system. According to the issues above, this paper proposes a personalized book recommendation model merged with personality trait, the main contents include the following aspects:Firstly, according to the definition of the "big five" personality trait model, this paper analyzes the personality trait factors, and measures its five dimensions, namely emotional stability、openness、extroversion、 agreeableness and conscientiousness. We calculate standard scores based on the mapping relationship between standard scores and original scores, then use strength relational tables between standard scores and personality, to measure user’s personality trait.Secondly, this paper proposes a personalized recommendation model merged with individual personality trait, including the factors of individual background factors drift, book type preference and personality trait. We preprocess the data to locate the target data set accurately by individual background factors drift, alleviate the sparseness of data by converting book scoring matrix to book type preference matrix, then optimize the sort of recommended book set by using personality trait compatibility, making users have a high satisfaction to the recommendation results. Thirdly, this paper comes up with evaluation indicator to measure the satisfaction of the recommendation results. According to the results of users’personality trait investigation, this paper confirms the size of recommended book set, then analyzes the results of the two recommendation algorithms, including users’satisfaction and average sort precision to indicate the advantage of individual recommendation algorithm merged with personality trait.