Science and technology input and output efficiencies have been a very important consideration in evaluating the innovativeness and evolution of traditional industries. In this research, the spatio-temporal trends of digital intelligence innovation performance in traditional industries during the provincial level in China between 2013 and 2022 are examined using the DEA-Malmquist index. Because of the existence of missing or ambiguous evaluation criteria, the paper presents a hybrid of fuzzy information granulation and support vector machine regression (FIG-SVM model) to address the voids in data. As it is shown by the results, the FIG-SVM model is successfully used to deal with fuzzy data and improves the assessment of the performance of innovation in traditional industries. In terms of time analysis, average values of total factor productivity, technical efficiency, and technological progress of the last decade are 0.969, 0.998, and 0.975 respectively, which indicates that there is great potential in improving the digital intelligence-based innovation in traditional industries. The geographical distribution of the performance of digital intelligence innovation in traditional industries depicts a trend of being stronger in the east, medium in the west, and weaker in the central regions.
