With the rapid development of science and technology, the big data market has emerged. In addition to the influence of the modern market environment, more and more enterprises pursue great profits, which requires enterprises to have higher standards for their products. Therefore, product innovation is an inevitable trend. However, product innovation is a comprehensive problem, involving a wide range. Both the characteristics of the product and the consumer should be taken into account. In the design method, it is necessary to pay attention to the interrelation with big data. Firstly, the theory of big data technology is expounded, including the framework and the principle of big data technology, and the classification of data analysis. Secondly, this research explains the process and method of product design, as well as the market-based research process, and proposes the route of using big data technology for product innovation. Finally, the Microsoft COCO: Common Objects in Context (MS-COCO) data set is taken as the research object, and ten groups of data are selected. Through the Back Propagation Neural Network (BPNN) model and neural style transfer (NTS) model, the results of the product innovation design method of big data technology are predicted. Two experiments were carried out, namely, the comparison analysis of the prediction accuracy of BPNN and NTS model; the error analysis of BPNN prediction and NTS model, and the final quality assessment of the two products. The final results reveal that: (1) the prediction accuracy of the BPNN is around 95.71%, and the prediction accuracy of the NTS model is about 85.7%. Therefore, the prediction accuracy of the BPNN is obviously higher than that of the NTS model in the product innovation design method. (2) The mean absolute error (MAE) of the prediction of the BPNN model is about 0.08, and the MAE of the prediction of the NTS model is about 0.11. Thus, the error of the BPNN model in the product innovation design method for product prediction is small. Furthermore, based on the BPNN model, the quality assessment of the two products is finally carried out, and the error rate is 0.11% and 0.12% respectively, which is very small. Thereby, the BPNN model has higher accuracy in predicting the product innovation design methods. The big data technology studied here is of great significance to the product innovation design method and has certain research significance for the intelligent and systematic design of products. Combined with the BPNN model, the accuracy of the prediction ability of product innovation can be guaranteed to a certain extent. This article is protected by copyright. All rights reserved.
Read full abstract