Software Quality in Use Characteristic Mining from Customer Reviews Warit Leopairote, Athasit Surarerks, Nakornthip Prompoon Department of Computer Engineering,

Software Quality in Use Characteristic Mining from

Customer ReviewsWarit Leopairote, Athasit Surarerks, Nakornthi

p PrompoonDepartment of Computer Engineering, Faculty

of Engineering

Outline

• I. INTRODUCTION

• II. BACKGROUND• A. Software Quality• B. Ontology Based Opinion Mining

• III. EVALUATION

INTRODUCTION

• Because of time and resource constraints, customers may not view all products offered or available at an e-commerce website

• opinion mining that extracts, analyses and aggregates information.

INTRODUCTION

• This approach focuses on analyzing the sentimental sentence that results in a positive or negative polarity judgment.

• To classify sentimental sentence to product attributes, ontology mapping is used

INTRODUCTION

• One of the widely accepted among software engineers in software quality model is ISO 9126 which presents one part of quality model, named quality in use

• effectiveness, productivity, safety and satisfaction characteristic

Outline

• I. INTRODUCTION

• II. BACKGROUND• A. Software Quality• B. Ontology Based Opinion Mining

• III. EVALUATION

A. Software Quality

• Our research focuses on ISO 9126 presented by the International Organization for Standardization

• This model contains two parts. The first part is internal and external quality.

• funcionality, reliability, usability, effeciency, maintainability and portability

• The second part is the quality in use.• effectiveness, productivity, safety and satisfaction

A. Software Quality

Outline

• I. INTRODUCTION

• II. BACKGROUND• A. Software Quality• B. Ontology Based Opinion Mining

• III. EVALUATION

B. Ontology Based Opinion Mining

• Technically, sentiment analysis and opinion mining approaches are used to extract and analyze information from product reviews.

• Ontology based opinion mining is a method for opinion mining. It can also be used to simplify such various aspects into polarity.

• ontology mining can be divided into two main parts, ontology mapping and polarity mining

B. Ontology Based Opinion Mining

B. Ontology Based Opinion Mining

• Ontology mapping part consists of two processes: ontology definition and ontology mapping.

• Ontology definition may be constructed based on product attribute, feature and quality.

• Ontology mapping aims to match a sentence or a section of a customer review to terminology defined in ontology

B. Ontology Based Opinion Mining

• In polarity mining part, many works on sentiment analysis classifies documents by their overall sentiment

• To construct classifier from machine learning approach

• A. Data Preparation Phase• B. Classifier Construction Phase• C. Reviews Analysis Phase• D. Quality in Use Score Calculation Phase

B. Ontology Based Opinion Mining

• A. Data Preparation Phase• Tokens of sentence• If-tagging• Past-tense-tagging• No-or-not-tagging

B. Ontology Based Opinion Mining

• B. Classifier Construction Phase• 1) Ontology construction part:• WordNet3.0

B. Ontology Based Opinion Mining

• 2) Rule construction part• Two rule-based classifiers are created in this part• Sentences that tagged with “if”, “past tense” or “no

or not” are filtered out.• Text-Miner Software Kit (TMSK) and the Rule Indu

ction Kit(RIKTEXT)

B. Ontology Based Opinion Mining

• C. Reviews Analysis Phase• 1) Mapping each sentence onto quality model by

ontology• 2) Classifying each sentence into positive or

negative sentence categorized by two rule-based classifiers

B. Ontology Based Opinion Mining

• D. Quality in Use Score Calculation Phase• 1) Calculating quality in use score in each review

• 2) Calculating quality in use score of software

III. EVALUATION

• A. Data Collection and Preparation• 500 reviews from 10 software. 3,002 sentences in

500 reviews are denoted as positive, neutral or negative opinions

III. EVALUATION

• B. Relation between quality in use score and rating star

III. EVALUATION

• C. Quality in use characteristic ontology mapping

III. EVALUATION

• D. Polarity sentence classification

END