CAJ | 학술논문

在高校图书馆场景下，受限于图书特征信息的稀少，很难运用基于内容的推荐算法。本文针对图书馆推荐场景提出了一种结合社会化标签的基于内容的推荐算法。通过将图书的社会化标签与实体图书相结合的方式来补充图书的特征项。对于社会化标签中存在的标签模糊问题，本文在信息熵思想的基础上提出了一种模糊标签识别方法。对于识别的模糊标签，本文采用标签去除的方式来处理，以增加标签信息的可靠性。最后通过基于内容的推荐算法来测试模糊标签去除前后推荐的准确率，以此来判定模糊标签识别方法的准确性。实验结果证明，该模糊标签识别方法具有较好的准确性，可以较好的识别模糊标签。
재고교도서관장경하，수한우도서특정신식적희소，흔난운용기우내용적추천산법。본문침대도서관추천장경제출료일충결합사회화표첨적기우내용적추천산법。통과장도서적사회화표첨여실체도서상결합적방식래보충도서적특정항。대우사회화표첨중존재적표첨모호문제，본문재신식적사상적기출상제출료일충모호표첨식별방법。대우식별적모호표첨，본문채용표첨거제적방식래처리，이증가표첨신식적가고성。최후통과기우내용적추천산법래측시모호표첨거제전후추천적준학솔，이차래판정모호표첨식별방법적준학성。실험결과증명，해모호표첨식별방법구유교호적준학성，가이교호적식별모호표첨。
In the university library scene, it is difficult to use content-based recommendation algorithm subject to rare book feature information. This paper proposes a content-based algorithm to use social tagging as part of the content. The other part of the content is the book feature information. To deal with the problem about fuzzy tag, this paper proposes a fuzzy tag identification method on the bases of information entropy. When identifying a fuzzy tag, this tag needed to be re-moved. Finally, run the content-based recommendation algorithm to test the accuracy before and after removal of the fuzzy label. The aim is to determine the accuracy of the fuzzy label recognition method. The experimental results show that fuzzy tag identification method has good accuracy and it can identify most fuzzy tags.