Jurnal Penelitian Pos dan Informatika

Postface

JPPI EJOURNAL — Fri, 31 Dec 2021 00:00:00 +0000

Desain Sistem Speech Recognition Penerjemah Bahasa Toraja Menggunakan Hidden Markov Model

Fri, 31 Dec 2021 00:00:00 +0000

Bahasa daerah toraja merupakan salah satu bahasa daerah yang ada di provinsi sulawesi selatan tepatnya di kabupaten Tana Toraja Indonesia. Indonesia memiliki beragam Bahasa daerah, Bahasa daerah merupakan salah satu ciri budaya suatu daerah khususnya Bahasa Toraja. Penelitian ini bertujuan untuk membuat sistem Speech Recognition pada penerjemah Bahasa daerah Toraja dengan menggunakan Hidden Markov Model. Hidden Markov Model merupakan metode klasifikasi yang menggunakan pembelajaran dari Baum-Welch, yang mengestimasi beberapa parameter untuk menemukan model terbaik yang menggambarkan training set. Hidden Markov adalah algoritma berbasis probabilistik yang digunakan untuk menentukan kelas kata terbaik untuk setiap kata dalam data uji. Hasil dari penelitian ini adalah sistem penerjemah bahasa daerah Toraja yang menggunakan metode Hidden Markov dapat membantu siapa saja yang belum mengetahui menggunakan dan arti Bahasa Indonesia dari Bahasa daerah Toraja. Berdasarkan pengujian Black Box untuk masing-masing fungsional pada sistem penerjemah bahasa daerah Toraja diperoleh hasil yang valid; demikian juga, pengujian logika menggunakan metode Hidden Markov bebas dari kesalahan logika. Kemudian berdasarkan pengujian kuesioner yang telah diperoleh dari 20 responden diperoleh hasil 80.2% sangat setuju jika sistem penerjemah tersebut bermanfaat.

Implementation of Backoff Algorithm Bidirectional Forwarding Detection (BFD) and MPLS VRF for Fast Recovery Mechanism End-to-End Multi-Circuit (E2E)

Hillman Akhyar Damanik, Merry Anggraeni — Fri, 31 Dec 2021 00:00:00 +0000

Implementing and modeling a multi-circuit backbone technology in a service provider environment, for a multi-circuit link method with a fast recovery and redundancy process, is one of the most useful and promising developments and trends in building a policy system for mapping packet paths for future systems generation. The use of streaming video conferencing, mobile user devices and mobility, the shift from TDM networks to IP based, cloud computing services, smart cities, and the Internet of Things (IoT), and Content Delivery Network (CDN) are the main generators of growth. Failure mechanism and fast recovery of link failure in connectivity tested using liveness with minimum link primary value intervals (10), secondary links (20) and tertiary links (30). The smaller the interval value in the policy specified in the preference, the rule will be used or take the recovery path link action first. The process that will be experienced when the main link fails be connected or active on the secondary link. The learning process of the back-off algorithm then exchanges update packages with interconnected neighbors. The next process is 10 seconds to receive traffic from a failed link. The failure transfer link to the process will not take time, can be interpreted as a 0-1 process, and there is no packet drop and loss on the link that performs the recovery process.

Ensemble Method Builds a Predictive by Integrating Several Models for Accurate Answer Prediction on Chatbot

Merry Anggraeni , Hillman Akhyar Damanik — Fri, 31 Dec 2021 00:00:00 +0000

One of the technological advances that has had a positive impact is the Chatbot. A chat service which is actually a feature that has been very often used by tech-savvy people. But the difference is, the ones who reply in the chat process are robots or virtual characters. The chatbot will provide answers to the questions given to it which is basically the chatbot assigned to understand what context the user (user) means, then reply to it with the appropriate context. However, each context has a different input, human language has a very flexible way so that it is often found inaccuracies in the prediction of answers given by Chatbot. This could be due to the incorrect choice of algorithms for the classification of the context or the lack of training data provided. To overcome this, this study will focus on strengthening the prediction of chatbot answers with the method of the machine learning algorithm (Machine Learning) ensemble 5 classification technique heterogeneously between basic classifiers and meta-algorithms and using maximum voting (Majority Vote) or Hard Voting on type ensemble. Classification is the process of finding a model or pattern that can describe and differentiate classes in a dataset. The goal is that the model can be used to predict objects with unknown class labels. It was found that the accuracy results were 86% for the data set with 6 classes, the avg Macro for each precision and recall was 92%, and the f1-score was 89%. Weighted avg for precision is 93% and each for recall and f1-score is 86%

The Role of the Information and Communication Technology Sector in the Indonesian Economy

Karina Maryana, Sartika Djamaluddin — Fri, 31 Dec 2021 00:00:00 +0000

The contribution of ICTs to the overall production system in the economy is considered as one of the factors for growth and development of the national economy. The purpose of this research is to investigate the contribution of the ICT sectors to the Indonesian economy through the role of the ICT sectors in encouraging the growth of other sectors with linkage analysis, multiplier analysis and causative matrix analysis. Furthermore, this study investigates the sources of growth in the ICT sectors by decomposing the sources of change in its output. In particular, the input-output approach is adopted to describe the production activities and diffusion of the ICT sectors in the Indonesian economy for the period 2000-2014. The results show that apart from having large output multipliers, the ICT Manufacturing sector also encourages exports. The results of the decomposition show that the influence of the previous technology coefficient has a significant contribution to the output of the ICT sectors. However, these values have diminished in recent periods so that the role of the Indonesian ICT sectors as a GPT sector has become less than optimal.

Analisis Performansi Baremetal Provisioning pada Openstack Platform Berbasis Remote Virtualisasi Menggunakan Layanan Ironic

Joey Leomanz Bartolomiussihosa, Agus Virgono, Ridha Muldina Negara — Fri, 31 Dec 2021 00:00:00 +0000

Penggunaan cloud computing sudah cukup banyak, bahkan banyak dipergunakan untuk melakukan pekerjaan tugas komputasi kinerja tinggi. Cloud data center mengandalkan teknologi virtualisasi untuk meningkatkan produktivitas dan mengurangi kompleksitas yang akan diberikan ke pengguna akhir dalam mengakses layanannya yang diberikan. Aplikasi yang berjalan dengan menggunakan teknologi virtualisasi tidak terhindarkan terjadinya penurunan performansi. Kebutuhan konsumen akan komputasi tinggi yang tidak dapat dilakukan pada teknologi virtualisasi dapat disolusikan dengan melakukan komputasi langsung di atas baremetal namun tetap pada kompleksitas yang sangat rendah. Openstack direpresentasikan sebagai platform opensource, populer sebagai Infrastructure as a Service (IaaS) cloud platform yang bisa diimplementasikan sebagai private cloud maupun publik cloud. Openstack sudah mendukung dalam melakukan teknik virtualisasi maupun mengakuisisi baremetal. Penelitian ini dilakukan uji performansi antara baremetal dan virtual mesin dengan melakukan beberapa pengujian. Pengujian performansi berdasarkan sumber daya infrastruktur dari komputasi awan tersebut seperti CPU processing time, network throughput TCP, disk I/O, memory test throughput, jitter, dan packet loss. Hasil pengujian yang didapat, performansi baremetal ironic lebih baik dari virtual mesin.

Design and Development Archipelago Architecture Learning Media using Qr Code Technology

Hanna Pratiwi, Yulia Wahyuningsih, Yohana Christela Oktaviani — Fri, 31 Dec 2021 00:00:00 +0000

In this digital era, almost all aspects of human life depend on technology, education, business, government, and even daily life. Along with the increasing advances in computer technology used for human interest, the science that supports it also develops, one of which is the field of Human and Computer Interaction often called HCI. Input-related methods are essential to consider in HCI; one example of input is Quick Response Code (QR-Code) technology. QR-Code technology has been widely applied in teaching and learning, such as attendance models and foreign language learning media. In this research, we also use QR-code technology to introduce the culture of the architecture, especially about traditional houses in Indonesia combined with motor training through coloring activities. It is hoped that children will have no problem writing as they get older and help develop creativity. The method used in this research is the Human and Computer Interaction (HCI) test method focused on the design of the user's system so the implementation of the system can be carried out efficiently by taking into user needs. Based on research that has been done, QR-Code technology can be used to support learning media.

Preface

JPPI EJOURNAL — Fri, 31 Dec 2021 00:00:00 +0000