The Weighted Values of Solar Evaporation’s Environment Factors Obtained by Machine Learning
Abstract
Enhancing the efficiency of solar evaporation is important for solar stills. In this study, the weighted values of environment factors (descriptors) on the efficiency of solar evaporation are obtained by using a machine learning algorithm, random forest. To verify the advancement between random forest and mathematical data analysis, two traditional methods, pair wise plots and Pearson correlation analysis, are conducted for comparison. Experimental data are obtained from around 100 articles since 2014. The results indicated that traditional methods failed at obtaining reasonable weighted values, while random forest is competent. It is found that thermal design is the most significant descriptors to obtain a high efficiency. The lack of complete dataset is the main challenge for more in-depth and comprehensive analysis. This work may promote the studies on solar evaporation and solar stills.
