note2
DESCRIPTION
Note 2 Review analysisTRANSCRIPT
Introduction
• To study the customer reviews for Samsung Galaxy Note 2.
• Analysis to be done on minimum 100 reviews.
• Review to be extracted from http://www.flipkart.com.
Word Cloud
• The most frequent words are battery , good, phone, best, even, good, great, like, screen.
• After finding association for all of these words .
Association Conclusions
• When user is talking about phone he is mainly pointing out performance , big screen, and it’s configuration which is good for it.
• User is also talking about features like playing games and watching movies.
• User are also calling this phone as awesome, unbelievable and they are saying it as the topmost phone in the market at that time.
• They are also impressed with the processor and comparing it with HTC.
• Talking about screen they are liking the amoled screen which is good for playing movies and surfing internet on it is good.
Continued….
• They are also impressed with battery which is good for heavy usage.
• They are also asking for Dual Sims adoption in the phone.
• Also asking for more slimmer body.
Ratings Analysis
• Rating are classified into three categories based on Stars given to product.
• Impressed User: Ratings = 5 Stars.
• Satisfied User: Ratings = 3 & 4 Stars
• Dissatisfied User: Ratings < 3 Stars.
• Out of 100 user : 74 user are Impressed and only 6 user are dissatisfied.
• Satisfaction rate is excellent for this phone due to it’s multiple features present on the phone like big screen, good processor etc.
• So, Lived up to the expectation.
Clustering
• Optimum number of Cluster = 3.
• Cluster 1 refers to the customers who are talking about various features of the phone like battery, screen, camera etc. These are also the customer who are satisfied with the phone.
Clustering
• Cluster 2 refers to customers are who are also interested in buying white color which may not be available due to heavy demand . These customer criticizing both the Note 2 and flipkart for their service i.e. delay of shipment. They are comparing it with other phones and also says
that are better.
• Cluster 3 refers to customers who are interested in update of the software in India of the latest android version.
SVM Classification Analysis
• From SVM Model we see can see that out of 100 , 88 support vector is correctly classified.
• Having no of classes = 3.
• Leveled as Impressed = 62 , Satisfied=20 and Dissatisfied= 6
SVM Classification Analysis
• Evaluating the most important variable for prediction with respect to weight from the taken 100 review into consideration.
• Most important variable are: smart ,camera ,service , galaxi, excel, and time.
• Least important variable are : Actual, person, annoy, Ear, hate , goodsound.
LSA
From below diagram we can infer that the words like good, battery, screen, features are the important words which are contributing more .
Sentiment Analysis
• After doing sentiment analysis by checking the polarity of the each review
• We infer almost same result as given in Rating ( in Star).
• Polarity for words in each user is also classified correctly.
• There are customer who are satisfied with the product but they did not give rating as 5 Stars as it can be more improved in some features.
Conclusion
• Most of the user are satisfied with the phone .
• Customers are talking about different features about the phone like camera, battery usage , screen etc.
• So , the phone can be recommended to others due to it’s overall good rating .