摘要
针对农业信息化综合服务平台信息过载问题,构建了个性化农业信息推荐模型,重点研究了用户类别兴趣向量、用户特征词喜好向量和文档特征向量,建立了农业专业词典和中英文停用词典;采用遗忘函数按时间对特征词的权重进行更新,并对用户类别兴趣进行更新,实现用户模型的更新;采用余弦相似度进行推荐度计算,提出了个性化服务推荐算法;通过对推荐信息的URL参数统计获知推荐效果,进一步对个性化推荐模型进行修正。结果表明,该模型可根据用户兴趣制定推荐,为用户提供有价值的信息,满足用户个性化需求。
The personalization recommendation model of agricultural information was constructed in view of information overload on agricultural information service platform. The model focused on three vectors including user category interest vector,user feature words preferences vector and document feature vector,established agricultural professional dictionary and English-Chinese disable dictionary;And then, using forgotten function the model updated weights of feature words, interest items of the user category and user model by time. The last,the model calculated recommended degrees by the cosine similarity,and proposed personalized service recommendation algorithm. Through the URL parameter statistics of recommended information,this model can inform recommendation effect and further correct the model. The results showed that the model can formulate recommendations based on user interest,provide valuable information, and meet the peisonalized needs of users.
出处
《湖北农业科学》
2015年第16期4052-4056,共5页
Hubei Agricultural Sciences
基金
山东省自主创新专项(2012CX90204)
关键词
内容过滤
个性化服务
农业信息
信息推荐
content filtering
personalized service
agricultural information
information recommendation
