| 000 | 00000nam c2200205 c 4500 | |
| 001 | 000045881646 | |
| 005 | 20160926171003 | |
| 007 | ta | |
| 008 | 160630s2016 ulkd bmAC 000c eng | |
| 040 | ▼a 211009 ▼c 211009 ▼d 211009 | |
| 041 | 0 | ▼a eng ▼b kor |
| 085 | 0 | ▼a 0510 ▼2 KDCP |
| 090 | ▼a 0510 ▼b 6YD36 ▼c 310 | |
| 100 | 1 | ▼a 박상민 |
| 245 | 1 0 | ▼a Sentiment root cause analysis based on relations among sentiment words / ▼d Sang-min Park |
| 260 | ▼a Seoul : ▼b Graduate School, Korea University, ▼c 2016 | |
| 300 | ▼a vii, 92장 : ▼b 도표 ; ▼c 26 cm | |
| 500 | ▼a 지도교수: 백두권 | |
| 502 | 1 | ▼a 학위논문(박사)-- ▼b 고려대학교 대학원: ▼c 컴퓨터·전파통신공학과, ▼d 2016. 8 |
| 504 | ▼a 참고문헌: 장 78-90 | |
| 530 | ▼a PDF 파일로도 이용가능; ▼c Requires PDF file reader(application/pdf) | |
| 653 | ▼a Root Cause Analysis ▼a Sentiment Analysis | |
| 776 | 0 | ▼t Sentiment root cause analysis based on relations among sentiment words ▼w (DCOLL211009)000000068500 |
| 900 | 1 0 | ▼a Park, Sang-min, ▼e 저 |
| 900 | 1 0 | ▼a 백두권, ▼e 지도교수 |
| 900 | 1 0 | ▼a Baik, Doo-kwon, ▼e 지도교수 |
| 945 | ▼a KLPA |
Electronic Information
| No. | Title | Service |
|---|---|---|
| 1 | Sentiment root cause analysis based on relations among sentiment words (51회 열람) |
View PDF Abstract Table of Contents |
Holdings Information
| No. | Location | Call Number | Accession No. | Availability | Due Date | Make a Reservation | Service |
|---|---|---|---|---|---|---|---|
| No. 1 | Location Science & Engineering Library/Stacks(Thesis)/ | Call Number 0510 6YD36 310 | Accession No. 123054359 | Availability Available | Due Date | Make a Reservation | Service |
Contents information
Abstract
Precise user-preference analysis is needed to generate a user profile for intelligent personal assistant. Aspect-level sentiment analysis can extract user preferences from product reviews but cannot explain the reasons for the user preferences because existing sentiment analyses only retrieve the polarity of a product feature. Further, it cannot consider the influence of sentiment words. Neutral sentiment words are not also utilized in sentiment analysis because these words do not affect the polarity calculation of the product feature. We propose a novel method that can analyze the root cause based on relations among words to extract the sentiment root cause. We use the fuzzy formal concept analysis to extend the feature-level hierarchy. A fuzzy cognitive map of the relations is employed to extract the root cause from the causes. The results show that we improved the accuracy of the sentiment cause and determination of the sentiment root-cause compared with the term frequency-based and sentiment score analyses.
Table of Contents
ABSTRACT i 1. Introduction 1 1.1 Research Motivation 1 1.2 Research Purpose 2 1.3 Research Taxonomy 6 1.4 Organization of the Dissertation 9 2. Related Work 10 2.1 User-Preference Analysis 10 2.2 Fuzzy Theory 11 2.3 Sentiment Analysis 14 2.4 Ontology-based Analysis 16 2.5 Root-Cause Analysis 17 3. User-Preference Analysis 19 3.1 Device-Oriented User Preference Analysis 19 3.1.1 Problem Statement 20 3.1.2 User-Centric Product Recommendation 24 3.2 User-Oriented User-Preference Analysis 25 3.2.1 Problem Statement 25 3.2.2 Personal Ontology-based Sentiment Analysis 30 3.3 Content-Oriented User-Preference Analysis 31 3.3.1 Problem Statement 31 3.3.2 Multimodal Generative Story Graph Analysis 32 4. Sentiment Root-Cause Analysis 34 4.1 Architecture 36 4.2 Sentiment Ontology based on FFCA 38 4.2.1 Factual and Sentiment Ontologies 38 4.2.2 Hierarchical Tree Generation using FFCA 41 4.3 Sentiment-Cause Ontology based on FCM-R 44 4.3.1 Sentiment-Cause Ontology 44 4.3.2 Sentiment Root-Cause Extraction 47 4.4 Implementation 52 5. Experimental Evaluation 55 5.1 Experiment Dataset 55 5.2 Sentiment-Cause Ontology 57 5.3 Quantitative Evaluation 65 5.3.1 Sentiment-Cause Analysis 65 5.3.2 Sentiment Root-Cause Analysis with Frequency 66 5.3.3 Sentiment Root-Cause Analysis with Sentiment Score 67 5.4 Qualitative Evaluation 68 5.5 Discussion 72 6. Conclusion and Future Works 76 6.1 Conclusion 76 6.2 Future Works 77 Bibliography 78 Abstract (Korean) 91