Research on Risk Assessment of Land Acquisition Based on Neural Networks Model

The application of artificial neural network technology in assessing land prices hinges on establishing the factor system that influences land prices. However, there is currently a lack of academic research on the factor system affecting rural land acquisition prices. To address this gap, our study utilizes a genetic algorithm to optimize the connection weights of neural networks, integrating a methodology that retains the best individual during the evolution process. This approach resolves the challenge of low network calculation accuracy caused by the initial weight randomness of neural networks, and mitigates the susceptibility to falling into local solutions. Considering the relatively limited transaction volume in the rural land acquisition market and the suitability of genetic neural network models, we propose using empirical methods and partial correlation analysis to determine the influencing factor system involved in model calculation. Genetic neural networks can effectively and objectively assess rural land acquisition prices based on market information, offering practical applicability.