CLUSTERING AND IDENTIFY GAMES REVIEW ON STEAM STORE

PRAYUDA, ADVERINO PUTRATAMA (2021) CLUSTERING AND IDENTIFY GAMES REVIEW ON STEAM STORE. Other thesis, Universitas Katholik Soegijapranata Semarang.

[img] Text
17.K1.0058-ADVERINO PUTRATAMA PRAYUDA-COVER_a.pdf

Download (673kB)
[img] Text
17.K1.0058-ADVERINO PUTRATAMA PRAYUDA-BAB I_a.pdf

Download (182kB)
[img] Text
17.K1.0058-ADVERINO PUTRATAMA PRAYUDA-BAB II_a.pdf
Restricted to Registered users only

Download (183kB)
[img] Text
17.K1.0058-ADVERINO PUTRATAMA PRAYUDA-BAB III_a.pdf

Download (292kB)
[img] Text
17.K1.0058-ADVERINO PUTRATAMA PRAYUDA-BAB IV_a.pdf

Download (559kB)
[img] Text
17.K1.0058-ADVERINO PUTRATAMA PRAYUDA-BAB V_a.pdf

Download (582kB)
[img] Text
17.K1.0058-ADVERINO PUTRATAMA PRAYUDA-BAB VI_a.pdf

Download (231kB)
[img] Text
17.K1.0058-ADVERINO PUTRATAMA PRAYUDA-DAPUS_a.pdf

Download (242kB)
[img] Text
17.K1.0058-ADVERINO PUTRATAMA PRAYUDA-LAMP_a.pdf

Download (420kB)

Abstract

Modern entertainment can already be achieved very easily, one example is a game that can be purchased online and downloaded immediately. As a means, online stores provide a review feature as a means to convince buyers through feedback from other buyers. However, this feature is often misused for one-sided advantage, including the Steam platform which has a large number of users. This research will discuss whether Steam has fake reviews and, whether each review is a helpful review. These questions will be solved by using a data mining algorithm. The use of K-Means will be focused on determining the majority of reviews are Recommended or Not Recommended reviews. The first result is big upvotes on each review mostly come from 2 factors, which is Steam Curators and big developers. The second result is long review and total play hour is not a determinant for the number of helpful votes. The third result is the accuracy from clustering Review and Recommendation using K-Means algorithm show a high percentage, with 500 sample data, this algorithm can have 100% accuracy, but with different amount of sample data the accuracy score can changed, in example if I use 100 sample data, the algorithm can only achieve 84% accuracy.

Item Type: Thesis (Other)
Subjects: 000 Computer Science, Information and General Works > 004 Data processing & computer science
Divisions: Faculty of Computer Science > Department of Informatics Engineering
Depositing User: mr AM. Pudja Adjie Sudoso
Date Deposited: 15 Oct 2021 03:12
Last Modified: 15 Oct 2021 03:12
URI: http://repository.unika.ac.id/id/eprint/27133

Actions (login required)

View Item View Item