a new approach and algorithm of sentiment analysis and product rating

Post on 14-Apr-2017

144 Views

Category:

Education

1 Downloads

Preview:

Click to see full reader

TRANSCRIPT

By- Nidhi Baranwal MCA-5th sem

Project involves

• Shopping portal for some specific product (Smart phones)• Make facility to comment• Analyze the comment• Do sentiment analysis using keywords• Rate the product automatically using algorithm

Approach for Sentiment Analysis

• Make a dictionary type database• Store keywords related to positive ,negative and neutral

sentiments• Search the proper words in the comments given by the user

similar to that stored in the database• Compare it with the keywords already stored in database to

find its polarity and corresponding polarity value

Algorithm for rating of the Product

1. Make a table having field ‘comment’, ‘point’, ‘temp’, ‘percent’, and ‘rate’ to store data

2. If “not rated” 2.1 search the keyword in comment and assign the point

correspondingly 2.2 convert it in percentage form and Put this value in ‘temp’

and ‘percent’ field 2.3 allot the rate value to ‘rate’ field on percentage basis3. Else 3.1 search the keyword in comment and assign the point correspondingly

Contd…

3.2 convert it in percentage form and take the previously stored value of ‘temp’ which is float value and already in percentage form

3.3 calculate the mean of both these values and assign it in ‘temp’ field3.4 now find out specific value for mean value within range and

convert it in percent form3.5 assign it in ‘percent’ field3.6 allot the rate value to rate field on percentage basis

Implementation

• Fetch the comment• Tokenize it• Remove irrelevant and noisy data• Feature extraction• Opinion word extraction• Opinion word polarity extraction• Summary generation• Scaling and rating

Technology used

• Front end- asp.net• Back end- microsoft sql database• Coding language- C#• IDE- visual studio 2010• Type- web application

Pictorial view

Scope

• This system is useful for the users who need reviews about a product• This system can be used with little investment and is

economically feasible• This application also works as an advertisement which makes

many people aware about the product• User can easily find out correct product for his usage

top related