A novel approach for evaluating the value of patent securitization basic asset pool based on machine learning

This paper aims to explore a method for evaluating the quality of underlying assets in patent securitization. We propose a new method to evaluate the quality of patent securitization, which combines the machine learning method with the income method to evaluate the quality of the underlying assets of patent securitization. First, we use random forest to predict the operating income of enterprises in the next few years; Secondly, according to the constructed patent value evaluation index system, we use random forest to determine the patent revenue sharing rate; Then, we calculate the patent value according to the present value formula; Finally, we verify the accuracy of the model according to the actual transaction data. The model we propose for evaluating the quality of assets underlying patent securitization demonstrates both feasibility and reliability. Firstly, the random forest method exhibits high accuracy in predicting operating income. Secondly, while the model does contain some errors compared to actual conditions, it overall performs well and can be effectively applied in the practical evaluation of the quality of underlying assets in patent securitization. The main contributions of the article are threefold: First, in the context of patent securitization, we study the problem of patent value evaluation, and supplement the existing literature. Secondly, based on the machine learning model, we put forward a new idea of the value evaluation of underlying assets in patent securitization, and effectively evaluated the economic value of patents. Finally, we provide a practical and effective model for the quality evaluation of the underlying assets of patent securitization. The accuracy meets the practical requirements and can be used in the practice of patent securitization.