Applied Deep Learning with Keras
Thoughdesigningneuralnetworksisasought-afterskill,itisnoteasytomaster.WithKeras,youcanapplycomplexmachinelearningalgorithmswithminimumcode.AppliedDeepLearningwithKerasstartsbytakingyouthroughthebasicsofmachinelearningandPythonallthewaytogaininganin-depthunderstandingofapplyingKerastodevelopefficientdeeplearningsolutions.Tohelpyougraspthedifferencebetweenmachineanddeeplearning,thebookguidesyouonhowtobuildalogisticregressionmodel,firstwithscikit-learnandthenwithKeras.YouwilldelveintoKerasanditsmanymodelsbycreatingpredictionmodelsforvariousreal-worldscenarios,suchasdiseasepredictionandcustomerchurning.You’llgainknowledgeonhowtoevaluate,optimize,andimproveyourmodelstoachievemaximuminformation.Next,you’lllearntoevaluateyourmodelbycross-validatingitusingKerasWrapperandscikit-learn.Followingthis,you’llproceedtounderstandhowtoapplyL1,L2,anddropoutregularizationtechniquestoimprovetheaccuracyofyourmodel.Tohelpmaintainaccuracy,you’llgettogripswithapplyingtechniquesincludingnullaccuracy,precision,andAUC-ROCscoretechniquesforfinetuningyourmodel.Bytheendofthisbook,youwillhavetheskillsyouneedtouseKeraswhenbuildinghigh-leveldeepneuralnetworks.
·7.6萬字