Mining and Analyzing Implicit Dataset of Domestic Load Consumption using Neural Network

The recent rise and fall of interest in conserving energy has generated an amazing quantity of load information being used, which boosts the data-driven algorithms for broad application throughout the building trade. This paper presents the recurrent neural network model to make short to medium team predictions using energy consumption data in residential buildings at one-hour resolution. This project reviews the prevailing data-driven approaches under different archetypes including methods for prediction and methods for classification. With advances in sensors and smart technologies, there is a need for short to long term prediction of electricity consumption in residential and commercial buildings to support decision making pertaining to operations, demand response strategies, and installation of distributed generation. Significantly, this project refines a few key tasks for modification of the data-driven approaches in the context of application of neural network to building energy analysis. The conclusions drawn in this project could facilitate future micro-scale changes of energy use for a particular building. For predicting the commercial building's load profiles, the proposed RNN models generally correspond to lower relative error. All these will be useful to establish a better long-term strategy for urban feasibility.