The International Journal of Engineering and Science (IJES) || Volume || 11 || Issue || 6 || Series I || Pages || PP 01-08 || 2022 || ISSN (e): 2319-1813 ISSN (p): 20-24-1805 DOI:10.9790/1813-1106010108 www.theijes.com Page 1 Machine Learning Application in Solid Waste Management: a review of Literature Titus Jefwa 1,2 , KennedyOndimu 2 ,KennedyHandullo 2 1. Coastal and Marine Resource Development(COMRED) | 2. Technical University of Mombasa --------------------------------------------------------ABSTRACT---------------------------------------------------------------- Inthispaper,wepresentacomprehensivereviewofresearchdedicatedtoapplicationsofmachinelearning in Solid waste management. The works analyzed were categorized in classes of threegenericcategories;namely,predictionofwastegenerationmodel,wastedetectionmodels,optimization of collection and disposal models. The paper reviewed studies from 2008 thatfocusingthethreedomainandthedifferentmachinelearningmodelsusedtosolvewastemanagementchallenge.The analysis prioritized domain in prediction of generation, detection and finally optimization of collection solid waste, the findings indicated GIS-based optimized using ArcGIS Network Analyst tool applied on variables such as cost, route distance and number of trucks, gives the best results. Further research will be carried out in future to realize and validate the tool. KeyWords:ArtificialIntelligence,MachineLearning,Modeling,Optimization,Deeplearning,Neuralnetwork --------------------------------------------------------------------------------------------------------------------------------------- Date of Submission: 15-06-2022 Date of Acceptance: 30-06-2022 --------------------------------------------------------------------------------------------------------------------------------------- I. Introduction Cities generate economic growth accompanied by solid waste depending on various urban forms (Lehmann,2011).TheproblemofSolidWasteManagement(SWM)ismultidimensionalandisbestappreciated in the light of rapid urbanization (Oteng-Ababio, 2010). Solid waste management(SWM) plays a critical role in the global economy. Pressure on management of the waste systemincreaseswiththecontinuingexpansionofthehumanpopulation.On the other hand, adaptingtonewtechnology improves the precision of collection from generation to disposal. Machine learning is one of the new scientific fieldsthat usesdata-intensive approaches leading to accurate and fasterdecisionmaking for effective management of waste. Machine learning is the core area of Artificial Intelligence (Dasgupta &Nath, 2016).Machine learning and artificial intelligence has merged with big data analytics has improvedperformance computing to create new opportunities for data intensive science in the multi-disciplinary domain (Liakoset al., 2018). It employs variety of statistical, probabilistic and optimization techniquesthatallowscomputersto“learn”frompastexamplesandtodetecthard-to-discernpatternsfromlarge,noisy or complex data sets (Cruz & Wishart, 2006). In this paper, we present a comprehensive review of the application of ML in solid wastemanagement (SWM). A number of relevant papers are presented that emphasize key and uniquefeatures of popular ML models. The paper is structured in sections as follows: Section 2 thedefinition of terms used in ML, most common models and algorithms. Section 3 presents theimplemented methodology for the collection and categorization of the presented works. Finally,in Section 4, the strengths derived from the implementation of ML in SWM are listed, as well asthefutureexpectations inthe domain. Machine learning is a branch of artificial intelligence research thatemploysavarietyofstatistical,probabilisticandoptimizationtoolsto“learn”frompastexamples and to then usethatpriortrainingtoclassifynewdata,identifynewpatternsorpredictnoveltrends(Mitchell, 1997). According to (Liakoset al., 2018), ML methodologies involve a learningprocesswith theobjectiveto learnfrom “experience”(trainingdata)to performatask. Machine learning, like statistics, is used to analyze and interpret data. Unlike statistics, though,machinelearningmethodscanemployBooleanlogic,absoluteconditionality,conditionalprobabilitiesandunco nventionaloptimizationstrategiestomodeldataorclassifypatterns(Cruz,