| 基于用户属性偏好与时间因子的服装推荐研究 |
| |
| 引用本文： | 周 静,何利力.基于用户属性偏好与时间因子的服装推荐研究[J].教育技术导刊,2020,19(6):23-28. |
| |
| 作者姓名： | 周 静 何利力 |
| |
| 作者单位： | 浙江理工大学 信息学院，浙江 杭州 310018 |
| |
| 基金项目： | 浙江省科技厅（重大）项目（2015C03001） |
| |
| 摘 要： | 针对服装推荐方法推荐精度不高、覆盖率低，不能充分挖掘用户潜在兴趣的问题，提出一种基于用户图像内容属性偏好与时间因子的服装推荐（UIACF）算法。通过构建深度卷积神经网络，提取服装图像中的服装属性，并据此形成用户属性向量，将基于用户属性偏好的相似度与基于时间因子的用户兴趣偏好相似度融合，构建用户偏好模型。将其与基于用户的协同过滤（UCF）算法、基于项目的协同过滤（ICF）算法及基于项目偏好的协同过滤（UCSVD）算法进行比较，结果显示，UIACF 算法准确率提高 14%。该算法为基于用户的服装协同过滤个性化推荐提供了一种新思路，用户潜在兴趣挖掘效率更高。
|
| 关 键 词： | 图像分类 用户偏好 协同过滤 服装推荐 时间因子 |
| 收稿时间： | 2019-09-27 |
Clothing Recommendation Research Based on User Attribute Preference and Time Factor |
| |
| ZHOU Jing,HE Li-li.Clothing Recommendation Research Based on User Attribute Preference and Time Factor[J].Introduction of Educational Technology,2020,19(6):23-28. |
| |
| Authors: | ZHOU Jing HE Li-li |
| |
| Institution: | School of Informatics and Electronics，Zhejiang Sci-Tech University，Hangzhou 310018，China |
| |
| Abstract: | Aiming at the problem that the clothing recommendation method has low recommendation accuracy and coverage，and can not fully tap the potential interest of users，we propose a clothing recommendation algorithm based on user image content attribute preference and time factor（UIACF）. By constructing a deep convolutional neural network，the clothing attributes in the clothing image are extracted， and the user attribute vector is formed accordingly，and the similarity based on the user attribute preference and the similarity based on the time factor-based user interest preference are fused to construct a user preference model. Compared with user-based collaborative filtering（UCF）algorithm，item-based collaborative filtering（ICF）algorithm and item-based collaborative filtering（UC? SVD） algorithm，the accuracy of UIACF algorithm is increased by 14%. The algorithm provides a new idea for personalized recommendation of collaborative filtering based on users’ clothing，and it is more efficient to mine users’potential interests |
| |
| Keywords: | image classification user preference collaborative filtering clothing recommendations time factor |
|
| 点击此处可从《教育技术导刊》浏览原始摘要信息 |
| 点击此处可从《教育技术导刊》下载免费的PDF全文 |
