Beez-Fm is a Singapore headquartered company specializing in BMS as a service (BaaS). The Building Management System (BMS) industry is involved in the design, development, and implementation of systems used to monitor and control building services, such as HVAC, lighting, security, and energy management.​ 

One of Beezfm key traits lies in its business model. Beezfm understands that their clients face hesitation to adopt BMS due to high intial capital expenditure, therefore beezfm installs the BMS for clients for free and only collect fees as a portion of the energy savings if they are realised.

However, it is difficult for Beezfm to prove that the energy savings observed by the client is achieved through the installed BMS and not due to external factors, leading to disagreements during the negotiation of fees.

Therefore, our solution to this problem is to help Beezfm develop an ML model capable of predicting energy consumption as if the Building Management System (BMS) was not installed, thus enabling Beezfm to clearly demonstrate the impact of the BMS on reducing energy usage to clients, facilitating effective fee negotiations.

In implementing our solution, we focused on three key aspects: Data Analysis, Feature Selection and Modelling.

Data Analysis helps us to formulate a comprehensive and consistent dataset through collecting and researching on important features for energy modelling.

Feature Selection helps us to select the most important features through utilising our domain knowledge and correlation matrices.

Modelling is a process where we choose appropriate machine learning models to train and to do prediction.