Litcius/Paper detail

Prediction Analysis using Random Forest Algorithms to Forecast the Air Pollution Level in a Particular Location

Puli Dilliswar Reddy, L. Rama Parvathy

20222022 3rd International Conference on Smart Electronics and Communication (ICOSEC)26 citationsDOI

Abstract

To forecast the degree of air pollution in a specific area of a region using techniques Innovative Random Forest against Naive Bayes. Two groups of algorithms are Random Forest and Naive Bayes. The technique was developed and tested on a 32516-record dataset. In a programming experiment, each approach was iterated N=5 times to identify different levels of air pollution. The threshold value is 0.05 percent, and the confidence interval is 95%. The G-power test is around 80% effective. When compared to Naive Bayes, the innovative Random Forest method (98.26%) offers higher accuracy (97.32%).Random forest has the highest accuracy in comparison to the Naive Bayes algorithm. Significance value for accuracy is 0.056(p>0.05), Precision 0.02(p<0.05) and recall 0.01(p<0.05) based on 2-tail analysis. Conclusion: Random Forest has improved performance when compared to Naive Bayes in the forecast of pollution in air.

Topics & Concepts

Random forestNaive Bayes classifierBayes' theoremAlgorithmIterated functionComputer scienceConfidence intervalStatisticsMachine learningMathematicsArtificial intelligenceBayesian probabilitySupport vector machineMathematical analysisAir Quality Monitoring and ForecastingTraffic Prediction and Management TechniquesVehicle emissions and performance
Prediction Analysis using Random Forest Algorithms to Forecast the Air Pollution Level in a Particular Location | Litcius