Research on enterprise radical innovation based on machine learning in big data background

With the advent of the big data era, some new quantitative tools for enterprise radical innovation research have also emerged. The use of algorithms in machine learning can improve the deficiencies of the current methods of evaluating the technological innovation capability of enterprises. Based on the characteristics of radical innovation and the evaluation mode of traditional innovation performance, an evaluation index system of enterprise radical innovation performance is constituted by six dimensions of resources, technology, product, management, commercial value and social value, and a performance evaluation model of enterprise radical innovation is established by applying the BP neural network. Through model training and simulation verification on radical innovation performance evaluation of the sample enterprises, the results show that the method has high reliability and the model has good generalization ability.