Analysis and classification for Software as a Service Reviews Using Naive Bayes

Abstract

with the quick development of cloud administrations, there has been a noteworthy increment in the quantity of online customer audits and sentiments on these administrations on various web-based social networking stages. These audits are a wellspring of important data as to cloud showcase position and cloud shopper fulfilment. This investigation investigates cloud purchasers' audits that mirror the client's involvement with Software as a Service (SaaS) applications. The surveys were gathered from various web-based interfaces, and around 4000 online audits were broke down utilizing feeling examination to distinguish the extremity of each survey, that is, regardless of whether the assessment being communicated is sure, negative, or impartial. Likewise, this examination builds up a model for anticipating the feeling of Software as a Service customers' audits utilizing a managed learning machine called a help vector machine. The opinion comes about demonstrate that 62% of the surveys are sure which shows that purchasers are in all likelihood happy with SaaS administrations. The outcomes demonstrate that the expectation exactness of the SVM-based Binary Occurrence approach (3-crease cross approval testing) is 92.30%, showing it performs better in deciding supposition contrasted and different methodologies (Term Occurrences, TFIDF). This work additionally gives important understanding into online SaaS surveys and offers the exploration group the main SaaS extremity dataset.