基于深度学习的推荐算法研究

发布时间：2018-05-29 00:09

本文选题：推荐算法 + 深度学习　；参考：《兰州大学》2017年硕士论文

【摘要】：随着互联网技术的快速发展,商品种类变得日益繁多,这导致用户不可能通过逐一浏览商品的信息的方式来快速挑选出最满意的或适合自己的商品,为提高选购效率,推荐算法应运而生,其作为一类与实际生活息息相关的机器学习算法,给用户带来便利的同时也给商家也带来了巨大的利益,可以称得上是"双赢"的算法.深度学习是近几年受到学者们广泛关注的一个研究领域,由于其深层架构的特性使得深度学习模型可以学习更复杂的结构,所以在语音识别,机器翻译,图像识别等领域深度学习算法均取得了令人瞩目的成果.本文将深度学习算法和推荐算法结合在一起进行研究,将深度学习中的门限循环单元模型(gated recurrent unit,GRU)应用于推荐算法中,并针对国外电影评分集—MovieLens数据集和自行爬取的国内电影评分数据集—豆瓣电影评分数据集应用该算法进行实际数据分析,将此算法与推荐领域其他主流算法进行比较,考察其优劣.
[Abstract]:With the rapid development of Internet technology, the variety of goods is becoming more and more diverse, which makes it impossible for users to quickly select the most satisfied or suitable products by browsing the information of the products one by one, in order to improve the efficiency of shopping. Recommendation algorithm emerges as the times require. As a kind of machine learning algorithm which is closely related to real life, it brings convenience to users and also brings huge benefits to merchants, which can be called "win-win" algorithm. In recent years, deep learning is a research field that has been widely concerned by scholars. Because of its deep structure characteristics, the deep learning model can learn more complex structure, so it can be used in speech recognition, machine translation. The depth learning algorithms in image recognition and other fields have achieved remarkable results. In this paper, the depth learning algorithm is combined with the recommendation algorithm, and the threshold recurrent unit model in depth learning is applied to the recommendation algorithm. This algorithm is used to analyze the actual data of the foreign film scoring set -Movie Lens dataset and the domestic film score data set which is crawled by ourselves. The algorithm is compared with other mainstream algorithms in the field of recommendation. Examine its merits and demerits.
【学位授予单位】：兰州大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：J943

【相似文献】