Garuda - Garba Rujukan Digital

Aplikasi Deteksi Jumlah Orang pada Area Indoor Untuk Mendukung Pelaksanaan PPKM dengan Metode YOLO Yoken Adinata; Kartika Gunadi; Indar Sugiarto
Jurnal Infra Vol 10, No 1 (2022)
Publisher : Universitas Kristen Petra

Show Abstract | Download Original | Original Source | Check in Google Scholar

Lately, the world is being hit by the Covid-19 outbreak which has caused several activities and sectors to be hampered, one of which is the buying and selling sector, both inside shopping centers and outside. Finally a solution emerged where each store was given a limit on the number of visitors to minimize the spread of the Covid-19 virus. However, there is a problem, namely monitoring the number of visitors is done manually so it is less effective and efficient.Because of the need for an application that can monitor the number of visitors when it has reached a certain limit the application will send a notification to the user. The method that used in this application to detecting the number of people is You Only Look Once (YOLO) The application has a feature so that the user can configure the parameters that will be used.Overall, the detection system used still has ambiguity in detecting small objects so that sometimes they are not detected as people. The rest, the system runs without problems from large to medium sized objects. On the other hand, overall survey respondents are satisfied with this system, this can be seen in the results of a survey taken from 15 respondents regarding the assessment of how this application helps the implementation of PPKM, and getting an average score of 8.53 out of 10, while the assessment of the ease of use of the application and user interface is 8.6 out of 10 and 8.4 out of 10.

Aplikasi Sistem Pakar Rekomendasi Makanan untuk Memenuhi Kecukupan Gizi William William; Kartika Gunadi; Anita Nathania Purbowo
Jurnal Infra Vol 10, No 1 (2022)
Publisher : Universitas Kristen Petra

Show Abstract | Download Original | Original Source | Check in Google Scholar

Nutrition problems are one of the common problems in Indonesia. The lack of knowledge about nutritional needs is one of the main causes of nutritional problems. To avoid nutrition problems, everyone needs to know their daily calorie needs so they don’t eat more or less food than the required calories. Eating more than the required calories can lead to obesity while eating less than the required calories can lead to malnutrition. This research was made to help reduce nutrition problems by creating an application that can help users to know their daily calorie needs and provide food recommendations that suit their caloric needs. By using the forward chaining method to collect personal data from user and the genetic algorithm method to filter food to be recommended.Based on the results of the program test conducted, the genetic algorithm method used was successful in filtering food and recommending it where the total calories from the recommended foods are close to the users total calorie needs although not perfect and sometime still exceed the calories needed but not in large amounts. The list of food menus used as a reference is a balanced diet according to Ministry of Health, but further researchers can use other references.

Ringkasan Ekstraktif Otomatis pada Berita Berbahasa Indonesia Menggunakan Metode BERT Franky Halim; Liliana Liliana; Kartika Gunadi
Jurnal Infra Vol 10, No 1 (2022)
Publisher : Universitas Kristen Petra

Show Abstract | Download Original | Original Source | Check in Google Scholar

In this modern era, information has become an important part of everyday life. In getting information several things can be done where one of them is by reading. With the increasing amount of information available on the internet, it is difficult for humans to keep abreast of developments. Online news is also one of the sources of information on the internet with a very large number and various topics. Reading the whole information sometimes also takes a long time. Therefore, it is necessary to make a summary of the available online news to reduce reading time and obtain relevant information. In this research, a summary of the news will be made by selecting important sentences from the news text. The method used in this research is Bidirectional Encoder Representations from Transformers with the addition of a transformer encoder layer.Based on the results of the tests that have been carried out, the pre-trained indolem/indobert-base-uncased model can produce the best F1-Score 57.17 for ROUGE-1, 51.27 for ROUGE-2, and 55.20 for ROUGE-L using abstractive reference and 84.46 for ROUGE-1, 83.21 for ROUGE-2, and 83.40 for ROUGE-L using extractive reference.

Aplikasi Pengukuran Tinggi dan Berat Badan Manusia Menggunakan Morphological Image Processing Nicky Nicky; Kartika Gunadi; Anita Nathania Purbowo
Jurnal Infra Vol 10, No 1 (2022)
Publisher : Universitas Kristen Petra

Show Abstract | Download Original | Original Source | Check in Google Scholar

Currently, there are still many measurements of height and weight body human done manually, where someone who wants to measure height and weight they must have a measuring device. This causes difficulties for people who do not have height and weight measuring facilities such as a stature meter to measure height and a scale to measure weight. But with the development of technology, especially in the image processing, this can be made easier. Estimates of human body height and weight can be known through taking pictures with an android smartphone camera, so it can make it easier to measure height and weight without using manual measurement tools.This study uses morphological image processing methods to recognize objects and determine the object's height from the image. To find out the weight of the object using the body surface area (BSA) formula. Both methods are combined into an applications.The results of the study the average accuracy of measuring height using morphological image processing was 98.6%. Meanwhile, the average accuracy of weight measurement using the body surface area formula is 80.7%.

Aplikasi Sentiment Analysis terhadap Trend Cryptocurrency pada Platform Twitter Menggunakan Library Textblob sebagai Alat Bantu Berinvestasi Ricky Chandra; Kartika Gunadi; Stephanus Antonius Ananda
Jurnal Infra Vol 10, No 2 (2022)
Publisher : Universitas Kristen Petra

Show Abstract | Download Original | Original Source | Check in Google Scholar

Along with the rapid development of information technology, many digital communication tools make it easier for people to access or share information. Twitter is one of the social media that has more than 1.3 billion users and more than 500 million tweets or tweets every day. The uniqueness of Twitter limits the number of writings to 280 characters, making Twitter a social media that contains sentiments about something. The cryptocurrency alone gets 4.1 million hashtag exposures per hour and has 2225 unique tweets per hour on the Twitter platform. The number of tweets related to Cryptocurrency causes investors to lose in terms of time because they have to manually assess a tweet. To overcome this, an effort that can be realized is to classify sentiments. One of the Natural Language Processing methods that have been developed for sentiment classification is TextBlob. In this thesis, an application will be made with sentiment analysis features using the Textblob Library, request tweets data using the Tweepy API, visualization of tweets data in the form of pie charts, tables, and word clouds, features that display the market price and history of cryptocurrency coins using the CoinGecko API and YFinance. as well as tweets from selected accounts. TextBlob Library testing is done by classifying results with 100 data that have been labeled by 2 examiners who have more than 1 year of experience investing in cryptocurrencies, the results obtained are 35% of the data have similarities between the results with the second tester, the application is tested with Tweets data request according to keywords, as well as application testing to display visualizations of Tweets data. A correlation test was conducted between the price change of cryptocurrency coins in 24 hours with the results of the classification of Tweet data and the Tweet volume of several coins. The conclusion that can be drawn from the correlation test is that when Tweet volume increases from the previous day, there will be a trend where the coin will increase or decrease. The results of the application web page where the application can display tweet data according to keywords and display visualizations, as well as display the price and history of the cryptocurrency market according to the available input.

Sistem Pakar untuk Mendiagnosa Kerusakan pada Sepeda Motor Kawasaki KLX 150 Menggunakan Metode Forward Chaining dan Certainty Factor Maria Eve Angeline; Djoni Haryadi Setiabudi; Kartika Gunadi
Jurnal Infra Vol 10, No 2 (2022)
Publisher : Universitas Kristen Petra

Show Abstract | Download Original | Original Source | Check in Google Scholar

The Kawasaki KLX 150 is an all-road or dual sport motorcycle, which means it can be used on the road or off-road. Nowadayas dirt bike riders are not just crossers, ordinary people are starting to like dirt bikes to be used as daily vehicles. Dirt bikes have engines and various kinds of devices or parts that can be damaged or be problematic. The damage that often occurs on a dirt bike is considered trivial and not understood. Therefore, an expert system was created that can detect damage to the Kawasaki KLX 150 motorcycle, with the hope that this research can help replace the role of mechanics to diagnose damage based on the symptoms experienced. The expert system to diagnose damage to the Kawasaki KLX 150 will use the Forward Chaining method and the Certainty Factor method. The use of forward chaining method in this expert system is to collect facts obtained from users so that the system will produce conclusions. The use of Certainty Factor in this study is to provide a level of confidence from the results of system diagnosis in the form of metrics. From this expert system, it can provide information about the name of the damage, how to handle it and the level of confidence in the diagnosis. Application testing for the diagnosis of damage to the Kawasaki KLX 150, using real data with experts, resulted in a system accuracy of 90%. The application for the diagnosis of damage to the Kawasaki KLX 150 is also considered complete, accurate, appropriate and easy to use (user friendly) by the user.

Sistem Presensi Mahasiswa Menggunakan Face Recognition Dengan Metode Facenet Pada Android Evelyn Evelyn; Rudy Adipranata; Kartika Gunadi
Jurnal Infra Vol 10, No 2 (2022)
Publisher : Universitas Kristen Petra

Show Abstract | Download Original | Original Source | Check in Google Scholar

As of today, the student attendance system in the University of Petra uses a QR code system to do their daily attendance. This QR code system has some flaws that are often exploited by the students. Students usually exploit this system by asking their peers to login into their accounts and take the attendance for them by scanning the shared QR code. Implementing face recognition can be one of the means of prevention. This attendance system is an Android based system using Facenet model for the facial recognition system. Formulas L2Norm and Cosine Similarity is used for comparison means for the results of the face recognition system. Results show that cosine similarity is most optimal when using the 0.5f threshold with the score of 0.5104218 and accuracy with the score of 0.77162087. Meanwhile, L2Norm results show that it is most optimal when using the 8.0f threshold with the score of 5.8973804 and accuracy with the score of 5.8973804.

Sistem Tracking Pengiriman Barang Pada PT. Afro Angkasa Express Dengan Menggunakan Framework Laravel 8 Michael Yohanes Wicaksono; Alexander Setiawan; Kartika Gunadi
Jurnal Infra Vol 10, No 2 (2022)
Publisher : Universitas Kristen Petra

Show Abstract | Download Original | Original Source | Check in Google Scholar

In this era of globalization, it is increasingly common for people to shop and sell their goods online. So that more and more shipping companies are there. With the increasing number of existing freight forwarding companies, PT Afro Angkasa Express must offer convenience and features to attract customers. The design of this goods delivery tracking system is intended to make it easier for PT Afro Angkasa Express customers to track goods. In addition, customers can place orders more easily through the website. Customers will also be provided with features such as on-site pick-up, so that PT Afro Angkasa Express will become one of the community's mainstays for goods delivery solutions. The Laravel 8 framework has advantages compared to other frameworks such as the Templating Engine feature that can design unique layouts and also reduce the impact of XSS (Cross Side Scripting) attacks.

Title

Found 8 Documents
Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title Search

Found 8 Documents Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title

Found 8 Documents
Search