Estimated Price Estimation for Building Renovation Projects in Chiang Mai University Using Artificial Neural Network Technique

Abstract

Chiang Mai University is located in Mueang Chiang Mai District, Chiang Mai Province. It was established in 1964 and has been in operation for over 60 years. Buildings within Chiang Mai University have been in use for a long time and have deteriorated due to extensive usage. The efficiency of these buildings has declined, resulting in the need for renovations, especially during the past 10 years.

For renovation projects, budget planning must be done in advance before implementation in each fiscal year. Rough cost estimation methods are commonly used, particularly the cost-per-unit-area estimation method. However, this method has been found to deviate from the planned budget. Currently, artificial intelligence technology has played an important role and serves as a tool to facilitate human data analysis and create predictive models.

This study aims to develop a cost estimation model for architectural renovation projects at Chiang Mai University using artificial neural network (ANN) techniques. A total of 45 project records, conducted between 2018 and 2024 with a minimum value of 500,000 baht, were collected and analyzed using RapidMiner software. The model's performance was evaluated using Root Mean Square Error (RMSE) and R-squared (R²) metrics. The results indicate that the ANN-based model achieved an R² value of 0.4311 (43.11% accuracy), outperforming the conventional cost-per-unit-area estimation method, which yielded an R² value of 0.2382 (23.82% accuracy). The ANN model provides more accurate preliminary cost predictions and is easy to implement. Furthermore, it can support budget planning and be adapted for use in other estimation tasks.