Research And Implementation Of Recommendation Algorithm Based On Association Rules And Text Categorization

With the development of data mining technology, text categorization, text extraction, text recommendation have gradually become a hot topic in network text mining research under the current situation of the generation and dissemination of network text.In order to solve these problems, the following research is done. (1) Two improved K-means clustering algorithms are proposed, which are sensitive to the initial clustering center and poor for the irregular clustering effect. For K-means algorithm, the clustering effect of irregular clusters is poor. By each iteration, the clustering center of each cluster is adjusted on each feature dimension to achieve more precise center adjustment. Experiments on datasets show that the proposed algorithm improves the classification accuracy compared with the other algorithms in F1-measure. (2) Text summarization based on keyword filtering and latent semantic association. Firstly, the article is segmented by using EM algorithm. Then the clustered blocks are clustered and clustered, and the synthetical method is used to select the sentences with higher weights and rankings as possible. The influence of different weighting schemes on the performance of text summarization is also studied. (3) Rule-based text recommendation method. A rule-based text recommendation method is proposed. The random sampling data are clustered, and the recommendation rules are proposed based on the potential rules and content similarity.The experimental results show that the proposed method has better diversity and recommendation accuracy than other methods. Finally, this paper proposes a news recommendation method based on the general clustering algorithm and the rule-based recommendation algorithm. It provides a personalized news recommendation algorithm, and provides targeted and diversified news to the users.