help

profilebcs
Mining_Customer_Requirement_from_Online_Reviews_Based_on_Multi-aspected_Sentiment_Analysis_and_KANO_Model.pdf

*

Abstract

Keywords-online reviews; customer requirement; sentiment analysis; KANO model

A. Feature Extraction

150

2020 16th Dahe Fortune China Forum and Chinese High-educational Management Annual Academic Conference (DFHMC)

978-1-6654-1518-7/20/$31.00 ©2020 IEEE DOI 10.1109/DFHMC52214.2020.00037

20 20

1 6t

h D

ah e

Fo rt

un e

Ch in

a Fo

ru m

a nd

C hi

ne se

H ig

h- ed

uc at

io na

l M an

ag em

en t A

nn ua

l A ca

de m

ic C

on fe

re nc

e (D

FH M

C) |

9 78

-1 -6

65 4-

15 18

-7 /2

0/ $3

1. 00

© 20

20 IE

EE |

D O

I: 10

.1 10

9/ D

FH M

C5 22

14 .2

02 0.

00 03

7

B. Sentiment Analysis

C. Customer Requirement Mining

151

A. Data Processing

B. Feature-sentiment joint extract

C. Sentiment analysis

152

C PMI

C

C negative PMI negative

C negative

S PMI negative

Score(w) C w positive

C neg

C C C

S S NS S

S D S NS D S

NS D S

D. KANO Conversion

pos iW

i negiW i

C i positivepos i C iW

C i negativeneg i C iW

C i positive i C(i,negative)

i C(i) i

posW pos

iW negW neg

iW

153

pos iW

neg iW

pos pos iW W

neg neg iW W

pos pos iW W

neg neg iW W

pos pos iW W

neg neg iW W

pos pos iW W

neg neg iW W

A. Sentiment Analysis Performance

micro_P micro_R micro_F1

TP micro P

TP FP

TP micro R

TP FN

micro P micro R micro_F1

micro P micro R

B. KANO Results

pos iW

neg iW

posW negW

posW negW

pos iW

neg iW

pos iW

neg iW

154

155

156