Research On Prediction Of Indoor Environment Based On Machine Learning

Environmental issues have become a focus of attention.At present,research and management of air quality are mainly concentrated outdoors.Since people spend most of their time indoors,the quality of indoor air has attracted wider attention in recent years,and its damage is no less than that of outdoor air.In this paper,the indoor environment quality data is taken as a sample,and based on machine learning,the prediction and optimization of the indoor environment is realized.The main work of this article is as follows:First of all,in view of the deficiencies in the application of air quality in traditional statistical and numerical models,this paper conducts research on the prediction and optimization of indoor air quality parameters based on machine learning,and validates the machine learning algorithm by analyzing the examples of machine learning algorithms.Actual application effect in air quality prediction and optimization.Secondly,in order to solve the problem of low sample utilization in the random forest algorithm,this paper improves and extends the algorithm to avoid the shortage of randomly sampled samples,and uses all samples to learn the base decision tree.At the same time,in order to ensure the independence of the base learner,a completely random selection method is used in feature selection.Experimental results show that the improved algorithm has obvious effects and reduces prediction errors to a certain extent.Finally,for the problem of unsatisfactory prediction results of a single model,the Boosting and Bagging algorithms only operate on training samples and connection methods without considering the performance of the base learner.Compared with the prediction evaluation and error comparison of other algorithms,it is found that the fusion regression prediction model has better prediction performance.