ANALYZING ICUSTOMER IREVIEWS: IA ITEXT ISUMMARIZATION ITECHNIQUE
Abstract
Reviews iare iunbiased iinformation iobtained ifrom ithe isources ioutside ian iorganization, which imakes ithem imore ireliable iin ithe ieyes iof icustomers. iOnline ishoppers iare ivery imuch iconcerned iabout iproduct ireviews ibefore imaking iany idecision iregarding ibuying ithe iproduct. iProduct ireviews iplays ian iimportant irole iin idetermining iwhat ikind iof iproduct iis. iSuch ireviews iprovide iuseful iinformation iabout icustomer iconcern iand itheir iexperience iwith ithe iproduct. iConsequently, ithese ireviews iwill ibe ihelpful ifor ia ibusiness imaking iproducts ifor ithe ipurpose iof iproduct irecommendation, ibetter icustomer iunderstanding iand iattracting imore iloyal icustomers.As iecommerce ihas ibecome iso ipopular, inumbers iof ireviews iare iincreasing iday iby iday. iIt iis idifficult ifor ia icustomer ito iread iall ithe ireviews imanually. iIn ithis ipaper, ian iapproach iis ideveloped iwhich iis iused ito iobtain ithe isummary ifrom ithousands ior ihundreds iof ionline ireviews. iThis iapproach iuses iextraction isummarization ifor isummarizing ithe ireviews ithereby iselecting ithe ioriginal isentences iand iputting iit itogether iinto ia inew ishorter itext iexplaining ithe ioverall iopinion iabout ithe iproduct. iAlthough iprevious istudies iof ideriving iuseful iinformation ifrom icustomer ireviews ifocus ion icategorical ior inumerical idata iand itextual idata ihas ibeen iignored. iBut itextual idata iare iof iequal iimportance iso iit ishould inot ibe iignored. iSo, ithis iapproach iincludes ievery iaspect iof ithe ireview iin ithe isummary iso ithat ia icustomer iwould iable ito imake ia iright idecision iregarding iproduct.