Abstract:
Occupant complaints are a reflection of poor building performance and an unsatisfactory indoor environment. One way to mitigate those complaints and to ensure occupants’ satisfaction with regards to building performance is through a well-performing facility management that is capable of planning for and addressing maintenance services. This thesis proposes a machine learning-based multistep generic framework to analyze occupant complaint data and to forecast the number of thermal complaints in particular for the upcoming week as part of the facility management’s predictive maintenance approach. Moreover, the developed forecasting model is benchmarked against a traditional statistical model to ensure proper performance. The proposed methodology was tested for a period of three years on a highly unstructured and unsolicited occupants’ complaints data recorded by facility management operators in a residential complex composed of 16 buildings. Text mining results of more than 6,000 occupant complaints showed that thermal related complaints are among the most common ones thus require further attention of facility managers. The developed Multi-Layer Perceptron (MLP) models to forecast the number of thermal complaints for the upcoming week showed proper performance with improvements over the traditional Autoregressive Integrated Moving Average (ARIMA) model with a higher ability to generalize to new data. It is also evident that the developed MLP forecasting models could assist facility managers in planning for the staffing resources required to handle these complaints thus enhancing occupant satisfaction and building performance.