Applied Machine Learning

Machinelearningmethodsarenowanimportanttoolforscientists,researchers,engineersandstudentsinawiderangeofareas.Thisbookiswrittenforpeoplewhowanttoadoptandusethemaintoolsofmachinelearning,butaren’tnecessarilygoingtowanttobemachinelearningresearchers.Intendedforstudentsinfinalyearundergraduateorfirstyeargraduatecomputerscienceprogramsinmachinelearning,thistextbookisamachinelearningtoolkit.AppliedMachineLearningcoversmanytopicsforpeoplewhowanttousemachinelearningprocessestogetthingsdone,withastrongemphasisonusingexistingtoolsandpackages,ratherthanwritingone’sowncode.Acompaniontotheauthor'sProbabilityandStatisticsforComputerScience,thisbookpicksupwheretheearlierbookleftoff(butalsosuppliesasummaryofprobabilitythatthereadercanuse).Emphasizingtheusefulnessofstandardmachineryfromappliedstatistics,thistextbookgivesanoverviewofthemajorappliedareasinlearning,includingcoverageof:•classificationusingstandardmachinery(naivebayes;nearestneighbor;SVM)•clusteringandvectorquantization(largelyasinPSCS)•PCA(largelyasinPSCS)•variantsofPCA(NIPALS;latentsemanticanalysis;canonicalcorrelationanalysis)•linearregression(largelyasinPSCS)•generalizedlinearmodelsincludinglogisticregression•modelselectionwithLasso,elasticnet•robustnessandm-estimators•MarkovchainsandHMM’s(largelyasinPSCS)•EMinfairlygorydetail;longexperienceteachingthissuggestsonedetailedexampleisrequired,whichstudentshate;butoncethey’vebeenthroughthat,thenextoneiseasy•simplegraphicalmodels(inthevariationalinferencesection)•classificationwithneuralnetworks,withaparticularemphasisonimageclassification•autoencodingwithneuralnetworks•structurelearning