IT110 project

Jack Bauer Task 2 -

Camera Recommendation

By Angela Vo

Alexandre Cortez

Nhat Le

Wen Zhong Wu

Objective

Jack Bauer is a field agent and asks for your decision support on cameras (Canon PowerShot SD500, Canon S100, Nikon Coolpix 4300 or Canon G3).

‹#›

Introduction

Sentiment Analysis

Sentiment is an attitude, thought, or judgment prompted by feeling.

Sentiment analysis is also known as opinion mining, studies people’s sentiments towards certain entities.

Sentiment analysis is one of the major tasks of NLP (Natural Language Processing).

Internet is a resourceful place with respect to sentiment information.

What and why useful?

Help us save time extracting big data and creating automation rating application.

Example:

Determine user ratings of products

Help develop recommender systems

Determine whether product reviews are fake

Disadvantage

Reviewer’s opinions might be subjective and make it hard to analyze.

Instead of sharing topic-related opinions, online spammers post spam

or irrelevant topic to product review.

Those things can make Sentiment Analysis inaccurate.

‹#›

Sentiment analysis method

A basic task in sentiment analysis is classifying the polarity of a given text at the document, sentence, or feature/aspect level

Review-level was chosen to examine the results for this project.

A number of online reviews about product was labeled as positive or negative on a scale of 5.

Using those reviews to create a training data to test which camera is best from given test data.

‹#›

Summary

How do we approach to get the data?

The data were manually collected on amazon.com and equivalent websites.

The requirements for our team is to collect 50 positive reviews and 50 negative reviews of each camera.

How do we combine the data?

With the data collected, we combine all positive reviews and negative reviews to be the training data.

Totaling 200 positive reviews and 200 negative reviews.

Close to 90% training and 10% testing data.

‹#›

Model

Process Documents Module - Process the documents

SVM Module (Support Vector Machine) - Supervise learning (Classified the two type of reviews positive and negative)

Apply Model Module - Apply the module to the testing data.

Process Documents Module - Process the documents

‹#›