Patterns Discovery onComplexDiagnosis andBiological DataUsingFuzzyLatent Variables

observations fromtheothers without knowledge ofany classlabels. Many clustering algorithms havebeen Thispaperproposes a newclustering algorithmproposed forbiological and medicaldata.These referred toasthePossibilitic Latent Variables (PLV) algorithms maybeclassified asbeing either hierarchical clustering algorithm. Thisalgorithm provides apowerfulclustering techniques orpartitional clustering methods. tool fortheanalysis ofcomplex data, suchasclinicalTheformer algorithms include agglomerative algorithms diagnosis andbiological expressions data, duetoits andgraphtheory clustering, andmergedataatdifferent robustness tovarious datadistributions andits accuracy levels toconstruct atree withbranches. Meanwhile, the inestablishing appropriate groupsfromdata.The latter category ofclustering algorithms, including those of algorithm combines adistribution modelandthefuzzy mixture decomposition, hardclustering anddeterministic degreesconcept. Comparedto theexpectation- annealing, divides datainto anumberofflat clusters. The maximization (EM)algorithm, whichisa well-knowngoalofbothclustering approaches istoidentify primitive distribution estimating algorithm, thePLValgorithm has groups inasetofobjects, points, orpatterns. Objects in theconsiderable advantage thatitcanbeapplied to onecluster aretightly coupled, and,whileobjects in various datatypes, i.e. itisnotrestricted solely to different clusters areloosely coupled. Gaussian datadistributions. Additionally, theproposed Traditional clustering algorithms assign eachobject algorithm hasabetter performance thanthewell-knownintoasingle cluster atthecompletion oftheclustering fuzzy clustering algorithm, i.e. theFCM algorithm, where task. Asanalternative tothese methods, fuzzy clustering itcanaddress compactregions, otherthansimply hasalsobeenshowntobe an effective clustering dividing objects intoseveral equalpopulations. The approach. Inthis method, eachdatapoint isassigned a performance oftheproposed algorithm isverified by fuzzymembership degreeindicating itsdegreeof conducting clustering tasks onthecontents ofseveralbelonging toeachcluster, rather thansimply being medical diagnosis andbiological expressions datasets. assigned toasingle cluster [3], [26]. Thismethodhad beenusedextensively andsuccessfully in computer