Intelligent interior design based on deep learning and CF algorithm

The application of recommendation algorithms in interior design helps users quickly determine their preferred design style and improve the efficiency. In response to the low personalisation and low recommendation accuracy in current home matching and style recommendation systems, research has been conducted on modelling interior design on the foundation of deep learning and collaborative filtering algorithms. Firstly, the collaborative filtering algorithm was optimised by combining content-based recommendation methods and principal component analysis. Meanwhile, a home matching recommendation model based on improved collaborative filtering algorithm was constructed. Then, a neural collaborative filtering style recommendation model based on channel attention was constructed by combining attention mechanism, convolutional neural network, and generalised matrix decomposition. These results confirmed that the improved home matching recommendation model had a higher accuracy of 88.4%. In the collected dataset of bedroom home matching schemes, the accuracy of the home matching recommendation model in home recommendations can reach 92.14%. The neural collaborative filtering style recommendation model on the foundation of channel attention had a hit rate of 0.8865. In summary, the constructed model has good application effects in interior design, which helps to promote the development of interior design.