As indicated by late examination, the measure of cell phone clients is bigger than PC clients. Simultaneously, the quantity of individuals who own Android telephones is expanding quickly. Android telephones bring individuals a ton of accommodation, in that it enables individuals to accomplish as much work as possible do on a PC, with no impediment by the area. Android has become a need instead of extravagance nowadays, and its prevalence has expanded quickly among accessible advanced mobile phones. There are bunches of OS which are accessible nowadays, yet among every one ofthem, Android is the best one, as it tends to be taken care of effectively and furthermore it is extremely simple to execute due to its open source nature. Android App Development has become a significant device for creating versatile applications. The Software Development Kit encouraged by Android helps engineers to begin creating and dealing with the applications momentarily, so the application can be actualized quicker. Since entrance testing is conceivable by utilizing the Android stage, there will be no compelling reason to convey your framework to different areas to do your pen test. As we as a whole know, entrance testing includes a lot of association ofthe individual into their framework, however by utilizing your Android telephone, you can perform it at any area in the most ideal manner you can.

Abstract

With the growth of ecommerce applications, almost everything is being sold and purchased online, which for the most part is quite significant. These customers kind of leave their reviews about purchased product which basically are then reviewed by other customers before making a purchase in a subtle way. By using Natural Language Processing on these reviews, we for the most part have proposed a supervised model which can give the polarity i.e., for all intents and purposes contrary to popular belief negativity and positivity of these reviews which can be used to specifically identify whether the product quality generally is good or not, for all intents and purposes. We have collected a data set ofsmartphones from amazon, it was then labelled to train and test the model on which we did comparative analysis of Logistic Regression, Linear Regression, Decision Tree, Random Forest, and Support Vector Machine classifier and other pre-processing methods and techniques which includes tokenization, lemmatization, stop words removal, and Parts of Speech Tagging to filter the data set. Term Frequency-Inverse Document Frequency was used to convert words to vectors for the training of model. In comparative analysis, we found best results of logistic regression algorithm. The accuracy we found against logistic regression was 86%. So, then used for further processing and predictions i.e., for logistic regression was predicting the quality of a product and popularity of a product. The data set of 5000 reviews of a smartphone was loaded for predictions which our model approximately predicted correctly.