Anomaly Detection for Vertical Plant Wall System using Novel Support Vector Machine in comparison with Linear Regression for improving accuracy

Aim: The purpose of the study and the goal is to create a plant wall system that can regulate the temperature and humidity of an interior environment by utilising a novel support vector machine in contrast to linear regression. Materials and the Methods: In order to identify the alterations in the indoor climate that are caused by vertical plant wall systems, a Machine Learning method for the regulation of indoor changes has been presented and developed. Results: In this method, Linear Regression and Innovative Support Vector Machine are compared with one another. After running the results of the calculation through the G Power (80%) calculator, the number 534 was decided upon as the appropriate size for the sample. 1068 different samples were used in total during this investigation. According to the findings, the Novel Support Vector Machine had the highest accuracy in forecasting changes inside of a building (68.42%), while the Linear Regression model had the lowest mean error (52.46%), and the value that was considered significant was 0.543. It illustrates that there is no statistically significant difference between the groups presented here. Conclusion: When it comes to accuracy in indoor control, the cutting-edge Support Vector Machine technique outperforms the more traditional Linear Regression method.