Skripsi

PERBANDINGAN TF-IDF DAN COUNT VECTORIZER TERHADAP KINERJA ALGORITMA MACHINE LEARNING DALAM ANALISIS SENTIMEN ULASAN APLIKASI DEEPSEEK

---

The rapid growth of artificial intelligence based applications has led to diverse user responses, which are commonly expressed through reviews on the Google Play Store. These user reviews not only reflect satisfaction levels but also provide valuable insights into user sentiment regarding application performance and quality. Therefore, sentiment analysis serves as an effective approach to systematically understand user perceptions. This study aims to analyze user sentiment toward the DeepSeek application by comparing two feature representation methods, namely Term Frequency–Inverse Document Frequency (TF-IDF) and Count Vectorizer, and evaluating the performance of several machine learning algorithms for sentiment classification. A total of 18,560 English-language reviews were collected through web scraping and labeled based on user ratings, resulting in 14,056 positive and 4,504 negative reviews. Feature representation was performed using TF-IDF and Count Vectorizer, followed by classification using Multinomial Naive Bayes, Support Vector Machine, Logistic Regression, K-Nearest Neighbors, and Random Forest. Model performance was evaluated using 10-Fold Stratified Cross Validation with accuracy, precision, recall, and F1-score metrics. The results indicate that TF-IDF generally outperforms Count Vectorizer. The combination of TF-IDF with Multinomial Naive Bayes and Logistic Regression demonstrated the best and equivalent performance, with both models achieving an accuracy of 87%, precision of 89%, recall of 95%, and an F1-score of 92%. These findings demonstrate that weighted feature representations combined with linear and probabilistic classifiers are effective for sentiment analysis of DeepSeek application reviews

---

Availability

Detail Information

Series Title

-

Call Number

T1950992026

Publisher

Indralaya : Prodi Sistem Informasi, Fakultas Ilmu Komputer Universitas Sriwijaya., 2026

Collation

xv, 102 hlm.; ilus.; 29 cm

Language

Indonesia

ISBN/ISSN

-

Classification

006.310 7

Content Type

Text

Media Type

unmediated

Carrier Type

-

Edition

-

Subject(s)

Prodi Sistem Informasi
Penambangan Teks (Text Mining)

Specific Detail Info

-

Statement of Responsibility

KA

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment