[目的/意義]傳統推薦方法僅考慮用戶過去的興趣偏好,忽略了用戶興趣偏好的漂移性問題,使得推薦結果過于專門化,不能給用戶提供新穎的推薦項目。[方法/過程]文章提出了一種基于語義關聯和信息距離的個性化推薦方法,該方法將項目的信息量融入到傳統的語義關聯相似度中,從而實現了用戶興趣偏好的有益遷移,使推薦得到有效擴展,改善了推薦專門化問題。[結果/結論]通過設計實驗驗證了信息距離能夠對推薦結果產生較大影響,提出的方法可以給用戶推薦其感興趣并且更有價值的項目。隨著新項目的不斷加入,項目的信息量會動態變化,系統會不斷調整推薦列表以適應用戶需求。[局限]不足之處在于模擬仿真實驗下樣本量不足引起的可信度問題,后續的研究將利用爬蟲工具收集大數據進行算法測試,驗證方法在大樣本環境下的有效性。

[Purpose/significance]Traditional recommendation methods only consider users’ past interests and preferences,ignoring the drift of users’ interests and preferences,which makes the recommendation results too specialized to provide users with novel recommendation items.[Method/process]This paper proposes a personalized recommendation method based on semantic association and information distance,which integrates the information content of items into the traditional semantic association similarity,thus realizing the beneficial transfer of user interest preference,effectively extending the recommendation,and improving the recommendation specialization.[Result/conclusion]Through design experiments,it is verified that the information distance can have a great impact on the recommendation results,and the proposed method can recommend more valuable items of interest to users.With the continuous addition of new items,the information content of the items will change dynamically,and the system will constantly adjust the recommendation list to meet the needs of users.[Limitations]The deficiency of this paper lies in the credibility problem caused by the insufficient sample size in the simulation experiment.The following research will use the crawler tool to collect big data for algorithm test and verify the effectiveness of the method in the large sample environment.