This book is based on the basis that machine learningalgorithms and rank based statistical methods are abetter choice to develop a robust model in logicalsituations. We designed experimental setup for datacollection, developed unique class of model includingvariable selection, and detection methods. Theselected significant variables provide a unique classof model for all six participants. We emphasize thebest selected variables have good information formodel development, and each selected variable have noerror i.e.; AUC=1, with forward selection and supportvector data description classifier. Basically, wedeveloped a unique class of model using six differentclasses of subjects, predicting elderly fallprevention, and after doing external validation withseventh class of subject, we reached a uniquesolution. Sections one is research introduction,section two is all about research design and dataanalysis, section three and four give extensivedevelopment of model for variable selection and oneclass classifier. Then finally given the conclusionand future aspect of whole study.