This study develops an automated click-tracking program to systematically access the official website of the English Premier League (EPL). Using artificial intelligence, the system performs deep text mining and data cleansing to extract specific textual information and convert it into structured data, which is then used to construct an AI-based vector database. Principal Component Analysis (PCA) is applied to identify key factors influenc-ing player salaries, while Data Envelopment Analysis (DEA) is employed to evaluate the performance and salary efficiency of EPL forwards by examining the relative relationship between input and output variables. To enhance reasoning and automate textual analysis, the open-source large language model TAIDE-LX-7B is fine-tuned to develop a generative AI capable of logical inference regarding player performance and salary rationality. The 2018/2019 EPL season is selected as a case study to validate the model’s accuracy and feasibility, and to compare its results with those proposed by [11]. Input indicators include height, age, salary, number of appearances, and shots taken. Output indicators include shots on target, shot accuracy, goals scored, goals per match, inverse average goal time, and number of wins. The model provides a comprehensive framework for assessing the performance of forward players and the fairness of their compensation.