ANALISIS SENTIMEN KOMENTAR INSTAGRAM TERHADAP POLEMIK DESAIN JERSEY TIMNAS INDONESIA DENGAN METODE ENSEMBLE

Abstrak

The controversy surrounding the design of the Indonesian national football team’s jersey has garnered significant attention on social media, particularly Instagram. Erspo, the brand responsible for the jersey’s design, is perceived to have fallen short of public expectations, sparking widespread discussions online. This study aims to analyze Instagram comments related to the jersey design using machine learning-based sentiment analysis techniques. Comments were collected through data crawling, followed by preprocessing steps such as text cleaning, tokenization, stemming, and stopword removal. Text features were transformed into numerical representations using the TF-IDF method, and sentiment classification was performed using individual models, including Random Forest, Naive Bayes, and Support Vector Machine. Furthermore, an ensemble model with VotingClassifier was applied to enhance sentiment classification accuracy. The findings reveal that the ensemble model achieved the highest accuracy of 77.4%, demonstrating its superiority over individual models. These results provide valuable insights into public perception of national product designs, offering a foundation for improving design and marketing strategies.

Polemik terkait desain jersey timnas sepak bola Indonesia menjadi sorotan di media sosial, khususnya Instagram. Erspo, merek yang bertanggung jawab atas desain jersey tersebut, dinilai belum memenuhi ekspektasi publik, yang memicu diskusi luas di media sosial. Penelitian ini bertujuan untuk menganalisis sentimen komentar di Instagram terkait desain jersey tersebut menggunakan pendekatan analisis sentimen berbasis pembelajaran mesin. Data komentar dikumpulkan melalui proses crawling, diikuti dengan langkah pra-pemrosesan, seperti pembersihan teks, tokenisasi, stemming, dan penghapusan kata umum. Fitur teks kemudian diubah menjadi representasi numerik menggunakan metode TF-IDF, yang dianalisis menggunakan model individu, yaitu Random Forest, Naive Bayes, dan Support Vector Machine. Selanjutnya, ensemble model dengan VotingClassifier diterapkan untuk meningkatkan akurasi prediksi sentimen. Hasil penelitian menunjukkan bahwa ensemble model memberikan akurasi tertinggi sebesar 77,4%, yang menunjukkan keunggulan metode ini dibandingkan model individu. Temuan ini memberikan wawasan penting mengenai persepsi publik terhadap desain produk nasional, yang dapat menjadi dasar perbaikan desain dan strategi pemasaran.