李智君

摘  要： 为了缩短室内设计推荐的时间，为用户快速推荐感兴趣的室内设计，提出基于协同过滤算法的室内设计推荐技术。通过计算室内设计分词，分析室内设计协同过滤的权重分布。基于对室内设计的内容进行关键词特征的选择和计算，确定室内设计特征提取流程，完成基于协同过滤的室内设计特征提取。根据用户的室内设计初始评分，计算出室内设计的权重，通过权重向量值预测用户室内设计的最终评分，利用室内设计推荐算法流程确定推荐算法的实现步骤，完成室内设计推荐算法的设计。最后通过室内设计协同过滤推荐模型，实现基于协同过滤算法的室内设计推荐。实验结果表明，基于协同过滤算法的推荐技术相比于传统推荐技术，室内设计的推荐时间缩短了70.3%。

关鍵词： 协同过滤算法; 室内设计; 推荐技术; 特征提取; 算法设计; 预测评分; 权重矩阵; 推荐模型

中图分类号： TN911.1?34; TP391                   文献标识码： A                    文章编号： 1004?373X（2020）13?0176?04

Interior design recommendation technology based on collaborative filtering algorithm

LI Zhijun

（Institute of Information Technology of GUET， Guilin 541001， China）

Abstract： An interior design recommendation technology based on the collaborative filtering algorithm is proposed to shorten the recommendation duration of interior design and quickly recommend interested interior designs for users. The weight distribution of collaborative filtering of interior design is analyzed by calculating the word segmentation of interior design. On the basis of the keyword feature selection and calculation for the content of interior design， the feature extraction process of interior design is determined and the interior design feature extraction based on collaborative filtering is completed. The weight of interior design is calculated according to the user′s initial score of interior design. The final score of user′s interior design is predicted by the value of weight vector. The flow of interior design recommendation algorithm is used to determine the implementation steps of the recommendation algorithm， so as to complete the design of interior design recommendation algorithm. In the end， the interior design recommendation based on collaborative filtering algorithm is realized by the collaborative filtering recommendation model of interior design. The experimental results show that the recommendation technology based on the collaborative filtering algorithm reduces the recommendation duration of indoor design by 70.3% in comparison with the traditional recommendation technology.

Keywords： collaborative filtering algorithm; interior design; recommendation technology; feature extraction; algorithm design; prediction score; weight matrix; recommendation model

0  引  言

科学技术的发展已经完全颠覆了人们的生活方式，人们日常生活的一些行为举动都逐渐由线下转至线上，网络的普及和电商的发展给人们带来方便的同时，也为人们塑造了一种极强的舒适感，无论身处何处，网络终端都可以为人们提供感兴趣的信息资源，还可以将人们的信息资源分享给外界[1]。如今用户的线上操作、资源共享、数据的产生导致网络数据量的增长，使人们不得不消耗大量的时间和精力去寻找更多有价值的信息，信息资源过载的现象也越来越严重。……