Pada bentuk soal essay atau uraian memberikan kesulitan tersendiri dalam mengoreksi karena jawaban essay bervariasi tergantung pada tingkat pemahaman terhadap suatu materi. Penelitian ini membuat sistem penilaian jawaban essay otomatis pada soal TIK menggunakan algoritma winnowing dan membandingkan dari penelitian sebelumnya yang menggunakan algoritma winnowing. Proses yang membedakan antara penelitian sebelumnya dengan penelitian yang diterapkan yaitu pada penelitian sebelumnya menggunakan metode winnowing dengan proses menghilangkan stopword removal dan proses stemming. Dari proses tersebut diduga waktu proses yang lebih lama sehingga proses yang akan diterapkan pada penelitian ini yaitu menghilangkan kata stopword removal dan menghilangkan kata yang dianggap mengulang soal dengan batas yang sudah ditentukan. Perbedaan proses tersebut sebagai dasar pengujian untuk mengetahui proses yang baik dan waktu yang lebih cepat. Hasil dari pengujian sistem menggunakan algoritma winnowig dengan perbedaan proses tersebut dengan hasil pengujian yaitu rata-rata nilai kemiripan 56.67% pada penelitian sebelumnya, sedangkan dengan penambahan proses pada pre-processing data dan penghilangkan proses stemming rata-rata kemiripan 55.504%. Perbedaan proses tersebut dianggap mendapat nilai rata-rata kemiripan yang sama. Sedangkan perbandingan waktu rata-rata berdasarkan proses tanpa stemming didapatkan 0.0077/s sedangkan pada peneliti sebelumnya yang melibatkan proses stemming mendapatkan hasil waktu 0.08/s. Kata kunci: Pre-processing, Stopword Removal, Stemming, Winnowing.

Teacher has different difficulties in correcting essay form because essay answers are various depend on the level of understanding towards certain material. This research makes assessment system of automatic essay answer for TIK question by using Winnowing algorithm and this research also compare to previous study which used the same method. The difference between this study and the previous study is the process. In previous study, the process used winnowing by omitting Stopword Removal and Stemming process. From the process, the researcher think that the process took longer time, thus the process that will be applied in this research by omitting Stopword Removal word and removing words which repeat the word in question based on the limit that had been already decided. This difference process becomes the basic testing to find out the better process and the faster time. The result of testing system by using Winnowing algorithm based on the difference process shows that the average of similarities value as much as 56.67% in previous study. Meanwhile, if the researcher adds the process to data Pre-Processing and omit the Stemming Process, the average of similarities values becomes 55.504%. By looking the difference process, the researcher thinks that the average of similarities is the same. While, if we compare to the average time based on the process without omitting Stemming, then the result becomes 0.0077/s. Meanwhile, in the previous study, the researcher involved the Stemming Process and the result was 0.08 %. Key word: Pre-processing, Stopword Removal, Stemming,Winnowing

Baca Jurnal Baca Jurnal Baca Jurnal Baca Jurnal Baca Jurnal Baca Jurnal