Skip to content

Latest commit

 

History

History
94 lines (55 loc) · 3.65 KB

README.md

File metadata and controls

94 lines (55 loc) · 3.65 KB
Raw

Sentiment-Analysis-on-Online-Product-Reviews

The objective of this project is to classify whether upcoming product will have positive or negative Sentiment.

alt text

Sentiment Analysis:

  • Sentiment analysis and Opinion mining is the computational study of User opinion to analyze the social, psychological, philosophical, behavior and perception of an individual person or a group of people about a product, policy, services and specific situations using Machine learning technique. Sentiment analysis is an important research area that identifies the people’s sentiment underlying a text and helps in decision making about the product.

Customer Product reviews:

  • A customer review is a review of a product or service made by a customer who has purchased and used, or had experience with the product or service. Customer reviews are a form of customer feedbacks on electronic commerce and online shopping sites. 90% of consumers read online reviews before purchasing product and 88% of consumers trust online reviews as much as personal recommendations.

Objective:

End-user:

E-Commerce Organization

Dataset:

The Online Product Reviews dataset includes following features: Dependent feature: reviews.rating (Unhappy,OK,Happy) Independent features: review.text (Reviews) The dataset contains 71044 rows and 25 columns

Workflow:

alt text

Steps:


1. Data loading.
2. Data Preprocessing

  • Remove Punctuations,special symbols and special characters.

  • Stopword Removal

  • Tokenization

  • Stemming

WordCloud:

alt text

Feature Extraction:

  • In order to make sense to our machine learning algorithm we have converted each review to a numeric representation which is called 'Vectorization'.

  • The system uses TF-IDF Vectorizer (Term Frequency-Inverse document frequency) that transforms a count matrix to a normalized frequency representation in float.

  • Splitting the data into Train and Test set (70-30).


1. TF-IDF

1.Count how many times does a word occur in each message (Known as term frequency) 2.Weigh the counts, so that frequent tokens get lower weight (inverse document frequency)


TF(t) = (Number of times term t appears in a document) / (Total number of terms in the document).
IDF(t) = log_e(Total number of documents / Number of documents with term t in it).

  • Example: Consider a document containing 100 words wherein the word dog appears 3 times. The term frequency for dog is (3 / 100) = 0.03. Assume total of 10 million documents and the word dog appears in one thousand of these. Then, the inverse document frequency is calculated as log(10,000,000 / 1,000) = 4.

  • Thus, the Tf-idf weight is the product of these quantities: 0.03 * 4 = 0.12.

Model Building:


Building Support Vector Machine (SVM) and SGDClassifier on feature vectors.

Results:

alt text


Thank You!

Follow @shreyaswankhede