Hanif, Irfan (2018) Klasifikasi Perintah Bahasa Natural Menggunakan Global Vectors for Word Representations (GloVe), Convolutional Neural Networks, dan Teknik Transfer Learning pada Aplikasi Chatbots. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
05111440000177-Undergraduate_Thesis.pdf - Accepted Version Download (2MB) | Preview |
Abstract
Akhir-akhir ini, chatbots menjadi salah satu alternatif yang sangat menjanjikan dalam menangani Frequently Asked Question (FAQ) yang diajukan konsumen-konsumen kepada organisasi atau perusahaan. Kini, ketidakmampuan chatbots dalam mengatasi keragaman kata yang digunakan para konsumennya menjadi masalah karena membuat mereka mulai frustasi dalam menggunakan chatbots tersebut. Maka dari itu, diusulkan metode klasifikasi perintah bahasa natural menggunakan Global Vectors for Word Representations (GloVe), Convolutional Neural Networks, dan teknik Transfer Learning pada aplikasi chatbots. Metode ini diusulkan karena kemampuannya yang bagus dalam membuat sistem chatbots mengenali makna kata, mengelompokkan kalimat-kalimat yang sering diajukan berdasarkan jawabannya, dan mengatasi permasalahan perusahaan yang tidak mempunyai data yang cukup untuk dipelajari chatbots. Ketiga algoritma ini terbukti memberikan hasil yang sangat baik untuk menjadi sistem inti chatbot karena hasil akhir uji coba model berhasil mencapai nilai F-Measure sebesar 95,6% dengan menggunakan data training yang relatif sedikit.
========================================================================================================
Lately, chatbots have become one of the most promising alternatives in dealing with Frequently Asked Questions (FAQs) that consumers propose to organizations or companies. Now, the inability of chatbots to cope with the diversity of words used by its customers becomes a problem because it makes them start frustating in using chatbots. Therefore, this final project proposed a natural language command sentence classification method using Global Vectors for Word Representations (GloVe), Convolutional Neural Networks, and Transfer Learning fot chatbots application. This method is proposed because its remarkable ability to make chatbots system recognize the meaning of the word, classifying FAQs sentences based on the answer, and overcoming the problem of companies that do not have enough data for chatbots to learn. These three algorithms proved to provide excellent results for being the core of chatbots system because the final test results of the model achieved F-Measure of 95.6% using relatively few training data.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Additional Information: | RSIf 004.019 Han k-1 3100018075488 |
| Uncontrolled Keywords: | chatbot; klasifikasi; GloVe; convolutional neural networks; transfer learning |
| Subjects: | Q Science > QA Mathematics > QA336 Artificial Intelligence Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science) Q Science > QA Mathematics > QA9.58 Algorithms |
| Divisions: | Faculty of Information and Communication Technology > Informatics > 55201-(S1) Undergraduate Thesis |
| Depositing User: | Irfan Hanif |
| Date Deposited: | 23 Jul 2018 04:55 |
| Last Modified: | 07 Oct 2020 01:56 |
| URI: | http://repository.its.ac.id/id/eprint/52194 |
Actions (login required)
 |
View Item |