Authors: Xiaoyuan Luo; Jun Liu
Addresses: Faculty of Science, Heihe University, Heihe Heilongjiang 164300, China ' Faculty of Science, Heihe University, Heihe Heilongjiang 164300, China
Abstract: In order to solve the problems of high recommendation error and long recommendation time in traditional personalised recommendation methods for smart home products, a new personalised recommendation method for smart home products based on convolution neural network is proposed. The attributes of smart home products are superimposed, and the square root of the attribute weight vector and all components are calculated. Determine the relationship between the attributes and important factors of smart home products to be recommended, and complete the weight calculation of smart home product recommendation. The personalised recommendation model of smart home products is constructed, and the convolution neural network is used to obtain the global optimal solution of the personalised recommendation model, so as to realise the personalised recommendation of smart home products. The experimental results show that the minimum error of the proposed method is about 0.3%, and the recommendation time is less than 15 s.
Keywords: smart home products; personalised recommendation; product attributes; convolution neural network.
International Journal of Product Development, 2022 Vol.26 No.1/2/3/4, pp.52 - 63
Received: 05 Feb 2021
Accepted: 08 May 2021
Published online: 07 Sep 2022 *