Optimal Recommendation Strategy Identification towards Energy Efficiency in Tertiary Buildings

Energy-related behaviour in tertiary environments is considered crucial when aiming towards energy efficiency in the long term, especially given the notion that people are not aware and actively engaged in their working place regarding energy consuming habits. Towards affecting this behaviour facility, managers engage various tactics, among which the provision of recommendations towards more eco-friendly energy-related behaviour. In an effort to identify the optimal persuasion strategy, this manuscript presents a socio-economic prediction model that evaluates data from an online survey (predictors) and identifies the optimal selection out of nine persuasion strategies (targets). A Random Forest Regression model was evaluated through over 350 responses, from 7 buildings across Europe before and after an ICT-based intervention with 9 persuasion strategies deployed, so as to evaluate the validity of the model. An assessment through Root Mean Square Error revealed average error rates below 1, in a scale from 1 to 5, which can be considered a promising result. Out of all principles evaluated, "Self Monitoring" and "Cause and Effect" presented the lowest error rates with RMSE 0.73 and 0.75, respectively.