Automated Essay Scoring menggunakan Naive Bayes dan Neuro Fuzzy

Muhammad Habibi

Informasi Umum

Kode

113090047

Klasifikasi

005.1 - Computer programming

Jenis

Karya Ilmiah - Skripsi (S1) - Reference

Subjek

Software Engineering

Dilihat

145 kali

Informasi Lainnya

Abstraksi

ABSTRAKSI: Natural Language Processing (NLP) merupakan bidang dari Ilmu Komputer dan Kecerdasan Buatan mengenai pemrosesan bahasa alami (manusia). Salah satu aplikasi NLP adalah Automated Essay Scoring (AES), yaitu sistem penilaian esai jawaban otomatis. Naïve Bayes sendiri bisa digabungkan dengan Neuro Fuzzy (Nefclass) karena Naïve Bayes mempunyai representasi data yang sama (nilai probabilitas term) dengan masukan sistem Nefclass (node-node di input layer) sehingga metode Naïve Bayes memungkinkan untuk bisa dipetakan. Hasil Tugas Akhir ini menunjukkan Naïve Bayes tidak cocok untuk digabungkan dengan Neuro Fuzzy jika diimplementasikan di studi kasus AES karena terjadi penurunan akurasi dan akan terdapat missing antecedent pada rule yang tercipta pada saat proses learning di Nefclass. Antecedent (sebab) yang hilang akan sangat berpengaruh terhadap consequent (akibat) yang didapat. Namun jika AES dikembangkan dengan Naïve Bayes saja dapat menghasilkan akurasi hingga 75%.Kata Kunci : Natural Language Processing, Automated Essay Scoring, Naïve Bayes, Neuro Fuzzy, Nefclass, antecedent, consequentABSTRACT: Natural Language Processing (NLP) is a field of Computer Science and Artificial Intelligence about human language processing. One of the applications of NLP is an Automated Essay Scoring (AES), which is an auto-scoring essay evaluation system. Naïve Bayes can be improved with Neuro Fuzzy Classification (Nefclass) because Naïve Bayes has the same data representation (the probability terms) as Nefclass system input (nodes in the input layer) so that the Naïve Bayes method makes it possible to be able to be mapped. The result of this final project is Naïve Bayes is not appropriate to improve it using Neuro Fuzzy if the study case is AES because there would be the degression of accuracy and missing antecedents in the rule that created in learning phase of Nefclass. Missing antecedents will affect to the consequent. But if the Automated Essay Scoring is developed using only Naïve Bayes could reach accuracy until 75%.Keyword: Natural Language Processing, Automated Essay Scoring, Naïve Bayes, Neuro Fuzzy, Nefclass, antecedent, consequent