@thesis{thesis,
		author={Ari Pradina  Yesika and Undari  Lilik},
		title ={THE QUALITY OF MACHINE TRANSLATION IN
TWITTER “ISLAMIC THINKING” ACCOUNT},
		year={2018},
		url={http://eprints.iain-surakarta.ac.id/2032/},
		abstract={Yesika Ari Pradina. 2017. The Quality of Machine Translation in Twitter
Islamic Thinking Account. Thesis. English Letters Study Program, Islamic
Education and Teacher Training Faculty.
Advisors : Hj. Lilik Untari, S.Pd. M.Hum
Key words : Machine Translation, Translation Quality Assessment,
Twitter, Islamic Thinking
The researcher analyze The Quality of Machine Translation in Twitter Islamic
thinking Account. In this research has three problem statement which are :
translation accuracy on machine translation in twitter, translation acceptability
on machine translation in twitter, and translation readadiability on machine
translation in twitter. The purpose of this research is to know what the result
of machine translation in twitter is accurate and acceptable.
In this research the researcher used descriptive qualitative method. The data
took from account twitter islamic thinking and observation. The other data are
taken questionnaries assessed by rater and respondents. The limitation  of the
data in this research are the tweet / the posting with periode february until
april 2016. The technique of data collection is documentation and
questionnaries. The technique of data analysis is collecting data, data
reduction, and data display.
In this research uses nababan theories about translation quality assessment.
The research findings 80 data screenshot. The dominant tweet in this
translation is complex sentences. In category accuracy the researcher conclude
there are 17 data (21.25%) as accurate, 30 data (37.75 %) as less accurate and
33 data (41.25 %) as inaccurate. . In category acceptability the researcher
conclude there are 20 data (25%) as acceptable, 29 data (36.25%) as less
acceptable and 31 data (38.75%) as unacceptable. In category readadiability
the researcher conclude there are 20 data (25%) as readiable, 29 data
(36.25%) as less readiable and 31 data (38.75%) as unreadiable.}
		}