Trích xuất câu gợi ý từ diễn đàn trực tuyến

Ngo Thi Lan Phirong vo Dtg Tap chf KHOA HOC & CONG NGHE 173(13): 67-72

TWCH XUAT CAU GOI Y TU" DIEN DAN TRT/C TUYEN

Ngo Thj Lan Phirong ,Ngd Thi Lan, Pham Thj Lien

Trudng Dai hpc Cdng Nghe Thdng Tin vd Truyen Thdng— DH Thai nguyen

TOM TAT

Cac phuang ti^n truyen thong xa hoi phat trien da tgo dieu kien thuan loi cho ngtroi dung chia se

quan diem, kinh nghiem va thao luSn cSc su kiSn trong th6 gidi thuc. C^c nSi dung thao liiSn, chia

se c6 nhieu thong tin httu ich, d^c biSt chiing se la cac ggi ^ cho nguoi dfing khde. Trong bai bao

n^y, chung tdi nghi€n ciiu phirong phap trich xuat tir dong cic c3u ggi y tren diln din true tuyen.

KSt qua thuc nghiSm fren bo du li?u tidng Viet frong mihn Imh vuc du lich, sii dyng mS hinh may

hgc Maximum Entropy la 77,68%. Kh qua du dk ap dung cho cac ung dung thirc tk nhu M thong

tu van, he thong hoi dap thong minh. Day la nghien ciiu dau tien ve phSn tich ggi y trong ngon ngi^

tieng Vi^t.

Tir khda: Khai phd gpiy, trich xudt gg-iy, phdn lich goiy, trich xudt ldi khuyen, dien ddn true tuyen

GIOI THIEU

Trong thai dai internet ngay nay, ngixai diing

CO ciing sa thich co the de dang chia se y

tucmg, thai do va y kien ciia minh tren cac

cong dong trirc tuyen (dien dan, nhom tren

mtmg xa hoi, website cho phep binh lu§n,

v.v). Han niia, hau het moi nguai thirong tgo

thanh cong dong de giup d& ian nhau trong

viec tim kiem, quang cao va ra quyet djnh. Do

do, CO rat nhieu thong tin hftu ich tren cac

cong dong true tuyen cho cac muc dich cu the

nhu: thuc phSm, du lich, siic khoe, tro choi,

phim anh, san pham, djch vu, v.v. Cac thanh

vien cua cpng dong c6 the trao doi kinh

nghiem tii cac trai nghiem thuc te ciia ho ve

mpt chii de nao do. Ngoai ra, cac thanh vien

khac CO th^ binh luan, nhan xet we kinh

nghiem vua diroc dua ra. Cu the se co mpt so

y kiSn d6ng y hoac co cac thong diep khuyen

cao va CO mpt so thong tin phan hoi sau khi

ho su dung kinh nghiem da dua ra do. Day la

cac thong tin gia tri, la cac go'i y t6t cho

nhiing ngudi diing khac dang tim kifim thong

tin hoac giai phap ve van de dang dupe thao

luan. D6ng thai, no cung la thong tin huu

dung, nhung gpi y t6t cho phep cac chu

thuang hi?u cai tien san pham/dich vu de

nang cao chat Itrpng san pham/dich vu ciia

minh. Mac dfi co nhung ggi y, nhutig kinh

' Tel: 0975 272359. Email: [email protected]

nghiem dupe chia se trSn internet nhung

thong tin thucmg tan mat, khong dl dk nguai

diing CO the tim kiem dupe cac thong tin gpi y

lien quan dfin van de ciia mjnh, thuang hp se

mat rat nhieu thai gian lang thang tren cac

diln dan khac nhau va se khong tong hpp het

duac cac ggi y phii hap. Do do, he thong co

th^ tu dpng phan tich thong tin gpi y (lai

khuyen, kinh nghiem, meo) tir cac y kien cua

nguai dung chia se tren dien dan de cung cap

cho nguai diing khac hoac chii thucmg hlfu co

thi tim kiem de dang la can thiet.

Kit qua trich xuat tu dpng cac cau gpi </ tu

diln dan co thi dupe tich hpp trong ling dpng

tim kilm thong minh va trong cac he thong tu

van nguoi diing. Van de phan tich gpi y dupe

xem nhu bai toan phan loai cau thanh hai

nhom: gpi y hoac khong phai gpi y.

Bai bao nay trinh bay phuang phap de trich

xuat cac cau gpi y trong cac bai viet cua dien

dan. Ve nguyen ly co the ap dung v6i cac

miln linh vuc khac nhau, nhung trong ph^m

vi nghien cuu cua minh chung toi lya chpn

ITnh vuc du lich, Trich xuat cac cau gai y cho

viec len ke hoac du lich ciia mpt khach hang

hoac cac cau ,gpi y giiip cho chu thuong hieu

cai tien chat lu'png san pham/dich vu cua hg.

Chung toi khao sat hieu qua ciia huong tiep

can may hpc (machine learning) su dung mo

hinh Maximum Entropy (MaxEnt) tren bg

ngir lieu (corpus) tieng Vi?t dugc xay dung tii

67