Auto-translate Youtube Pada Kanal AJ+ Sahah: Analisis Kesalahan Berdasarkan Penilaian TAUS dan Taksonomi Costa

Khasan Maulani Kiromim Baroroh, Dra. Uswatun Hasanah, M.A

2025 | Skripsi | SASTRA ARAB

Penelitian ini menganalisis kesalahan Auto-translate YouTube (ATY) dalam menerjemahkan takarir Arab-Indonesia pada domain puisi dan sejarah dari kanal YouTube AJ+ S??ah. Kualitas terjemahan dievaluasi berdasarkan penilaian TAUS (adequacy dan fluency, skala 1-4). Kesalahan diidentifikasi dan diklasifikasikan menggunakan taksonomi Costa dkk. Analisis data dilakukan dengan metode kuantitatif-kualitatif yang meliputi perhitungan frekuensi kemunculan kesalahan dan skor rata-rata TAUS secara kuantitatif serta klasifikasi kesalahan secara kualitatif.

Hasil penilaian TAUS menunjukkan skor adequacy 2,61 dan fluency 2,79 untuk domain sejarah, serta adequacy 2,52 dan fluency 2,81 untuk domain puisi. ATY cenderung menghasilkan keakuratan adequacy yang lebih tinggi dibandingkan skor fluency. Pembulatan mean fluency dan adequacy pada dua domain masing-masing memperoleh nilai 3, yang mengindikasikan bahwa mayoritas makna BSu berhasil diterjemahkan ke BSa meskipun terdapat masalah dari segi adequacy dan fluency.

Dari 100 segmen per domain, frekuensi kesalahan lebih tinggi pada puisi (116 kesalahan, 116%) daripada sejarah (80 kesalahan, 80%). Frekuensi kesalahan berdasarkan taksonomi Costa menunjukkan pola yang berbeda pada tiap domain: kesalahan semantis paling dominan pada puisi (38.8%), sedangkan kesalahan leksikal paling sering muncul pada sejarah (35.0%). Kesalahan tata bahasa sama signifikannya dalam kedua teks (~24-26%), sementara kesalahan ortografis dan wacana memperoleh presentase minimal (<7>

This study analyzes the errors of Auto-translate YouTube (ATY) in translating Arabic-Indonesian subtitles in the poetry and history domains of the AJ+ S??ah YouTube channel. Data analysis was carried out using quantitative-qualitative methods; quantitative calculation of error frequency and TAUS average scores (adequacy and fluency, scale 1-4); qualitative error classification based on Costa proposed taxonomy.

The TAUS assessment results showed an adequacy score of 2,61 and a fluency of 2,79 for the history domain, and an adequacy of 2,52 and a fluency of 2,81 for the poetry domain. ATY tends to produce higher adequacy accuracy than fluency scores. Mean rounding in both domain in fluency and adequacy each obtained a value of 3; indication that the majority of the SL meaning was successfully translated into the TL, though there were problems in terms of meaning and fluency.

Out of the 100 segments per domain, the error frequency was higher in poetry (116 errors, 116%) than history (80 errors, 80%). The error frequency based on Costa's taxonomy showed different patterns in each domain: semantic errors were most dominant in poetry (38.8%), while lexical errors were most frequent in history (35.0%). Grammatical errors were equally significant in both texts (~24-26%), while orthographic and discourse errors obtained minimal percentages (<7>

Kata Kunci : auto-translate YouTube, analisis kesalahan, TAUS, taksonomi kesalahan Costa