Final Year Project On Sentimental Analysis of User Feedbacks

Project Title: CLASSIFICATION OF SENTIMENT REVIEWS USING N-GRAM MACHINE LEARNING ALGORITHM Correction in Abstract: Along with Maximum Entropy also added Naive Bayes Description: When the Comment rating is usually used for feedback, people mostly don't spend their time in typing, so they just comment it as some neutral or above neutral comment even though they feel very good. This causes some inaccuracy while polarity analysis. But, people feel most of the time very easy to give a star rating coz its just a drag and click. The work here is to added or compare both the star and comment and produce a more accurate polarity. Eg: On a scale of very negative ,negative ,neutral ,positive and very positive, the polarity of comment rating is defined for a product. By the same way, the star rating is also rated on scale of 1-5 . now if the comment rating is positive and star is 1 , then the final polarity shall be around neutral. Likewise, the combination of both ratings are made and then the final result is produced for every reviews of product The interface should provide the query text and the search is made (eg: samsung s3 or macbook pro) for it in the dataset. the queried word matched with the available reviews and the number of positive or negative reviews are displayed in a Output textbox. an another output box for displaying the particular reviews that are changed. If a review shows as positive in comment rating and after comparing star ,if is shows as [login to view URL] the changed reviews alone has to be displayed in that output [login to view URL],use a graph or chart for displaying the percentage thats how i can show the panel member the accuracy of the reviews and objective of my project. I have attached the code i have already done( actually cloned from github and modified). Please, go through the code, understand how its done. use your skills to it. and finish it as per the requested output. i don't have the dataset,[login to view URL] of 800 reviews is [login to view URL] for single product domain or multiple. The code that has done is for twitter data and has some [login to view URL] files using twitter API. remove it and replace with product reviews, thats how it makes sense for adding star rating.I have not tested if it works with API independent of [login to view URL] your wish,if it works keep it or take [login to view URL] than that everything is relevant to project. there will be a [login to view URL] file, running that opens the GUI .The code works from [login to view URL] know what to do once you go through the code. Other Details: The [login to view URL] has [login to view URL] - the evaluator result should be Displayed in GUI. [login to view URL] - this evaluator result should be Displayed in GUI. [login to view URL] - it contains the confusion matrix and print statement that is being called by above two Evaluator files. [login to view URL], [login to view URL], [login to view URL] - does the classification process 5. [login to view URL] - does the feature extraction 6.customUI- run this first. I have no other detailed idea about the [login to view URL] have to go through the workings.I need this to be completed in 8-10 days. I hope it a simple process once you understand the code and project. I have added the GUI screenshot that display and a piechart that pops once the queried and check button is clicked. i want an additional box to it(bottom ,left or right) which displays the changed reviews after combining and comparing the both star and [login to view URL] box for showing the comment rating results and other shows the changed reviews after combining. Software i used: Pycharm and python2.7