Enhancing Zero-Carbon Building Operation and Maintenance: A Correlation-Based Data Mining Approach for Database Analysis

Abstract

In the context of global climate change and the increasing focus on carbon emissions, carbon emission research has become a prominent area of study. However, research in this field inevitably involves extensive monitoring, and when the data become complex and chaotic, the accuracy of these data can be challenging to control, making it difficult to determine their reliability. This article starts by exploring the operational and maintenance data of zero-carbon buildings, aiming to uncover the correlation between energy consumption data and environmental data. This correlation is categorized into two main types: linear correlation and trend correlation. By establishing error degree calculations based on these correlation relationships, anomaly detection can be performed on the data. Analyzing the interrelationships between these datasets allows for the formulation of appropriate fitting equations, primarily consisting of linear and polynomial fits, all of which exhibit a determination coefficient exceeding 0.99. These fitting equations are then utilized to correct errors in the anomalous data, and the reasonableness of the fitting methods is demonstrated by examining the residual distribution. The final results align with the corresponding expectations, providing a concise and effective correction method for monitoring data in zero-carbon smart buildings. Importantly, this method exhibits a certain level of generality and can be applied to various scenarios within the realm of zero-carbon buildings.