Eksplorasi Deep Learning DeepLDA Dalam Klasifikasi Judul Penelitian

Farandy, Bastian (2021) Eksplorasi Deep Learning DeepLDA Dalam Klasifikasi Judul Penelitian. Project Report. [s.n.]. (Unpublished)

[thumbnail of 05111740000190-Project_Report.pdf] Text
05111740000190-Project_Report.pdf - Accepted Version
Restricted to Repository staff only

Download (903kB) | Request a copy

Abstract

Penelitian merupakan salah satu bagian yang berperan penting dalam kehidupan manusia. Hasil penelitian menjadi salah satu faktor dalam mengantarkan manusia ke era modern seperti sekarang. Dengan jumlah penelitian yang semakin besar, maka diperlukan suatu teknik dalam melakukan klasifikasi judul penelitian tersebut.
Latent Dirichlet Allocation (LDA) adalah model probabilistik generatif dari koleksi data diskrit seperti korpus teks. Ide dasar dari LDA adalah merepresentasikan dokumen sebagai beberapa topik[1]. Proses LDA bersifat generatif melalui imaginary random process pada model yang mengasumsikan bahwa dokumen berasal dari topik tertentu. Setiap topik terdiri dari distribusi kata-kata[2].
DeepLDA merupakan sebuah algoritma yang menerapkan Latent Dirichlet Allocation (LDA) dalam deep learning. DeepLDA memanfaatkan LDA sebagai supervisor ketika melakukan training dari deep neural network (DNN), sehingga DNN dapat mendekati performa dari LDA dengan komputasi yang lebih rendah.
Untuk memenuhi klasifikasi dari judul penelitian, maka eksplorasi metode deep learning untuk klasifikasi topik ini dilakukan, salah satu nya melakukan eksplorasi dari metode DeepLDA.
========================================================================================================
Research is one part that plays an important role in human life. The results of research are one of the factors in bringing humans into the modern era like now. With the increasing number of studies, a technique is needed in classifying the research title.
Latent Dirichlet Allocation (LDA) is a generative probabilistic model of discrete data collections such as a text corpus. The basic idea of ​​LDA is to represent documents as several topics [1]. The LDA process is generative through an imaginary random process on a model that assumes that the document comes from a particular topic. Each topic consists of a distribution of words [2].
DeepLDA is an algorithm that implements Latent Dirichlet Allocation (LDA) in deep learning. DeepLDA uses LDA as a supervisor when training from a deep neural network (DNN), so that DNN can approach the performance of LDA with lower computation.
To fulfill the classification of the research title, an exploration of the deep learning method for this topic classification was carried out, one of which was to explore the DeepLDA method.

Item Type: Monograph (Project Report)
Uncontrolled Keywords: Penelitian, Latent Dirichlet Allocation (LDA), DeepLDA Research, probabilistik
Subjects: Q Science > QA Mathematics > QA76.6 Computer programming.
Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Q Science > QA Mathematics > QA76.9 Computer algorithms. Virtual Reality. Computer simulation.
Z Bibliography. Library Science. Information Resources > Z665 Library Science. Information Science
Divisions: Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User: Bastian Farandy
Date Deposited: 13 Jan 2021 04:41
Last Modified: 03 Apr 2023 06:48
URI: http://repository.its.ac.id/id/eprint/82398

Actions (login required)

View Item View Item