Pub Date : 2022-03-04DOI: 10.23917/khif.v8i1.12760
E. Ramadhani, Amrullah Sidiq
- Handling digital evidence in forensics is a very crucial task. Incorrect handling can cause the evidence to become invalid as proof of a crime in court. The procedure of handling digital evidence, starting from its collection, usage, and storage, affects its acceptability in the judicial process. Therefore, a digital evidence management system becomes imperative for police researchers and investigators. This study aims at designing such a system using the design thinking method, which goes through five stages: empathy, definition, idea, prototype, and test. The result of the study is a web-based system prototype. The prototype user testing attains a system usability scale (SUS) value of 60. The SUS value means that the prototype is in the category of marginal low and indicates that the prototype does not meet the feasibility and needs improvement.
{"title":"Design Thinking Method to Develop a Digital Evidence Handling Management Application","authors":"E. Ramadhani, Amrullah Sidiq","doi":"10.23917/khif.v8i1.12760","DOIUrl":"https://doi.org/10.23917/khif.v8i1.12760","url":null,"abstract":"- Handling digital evidence in forensics is a very crucial task. Incorrect handling can cause the evidence to become invalid as proof of a crime in court. The procedure of handling digital evidence, starting from its collection, usage, and storage, affects its acceptability in the judicial process. Therefore, a digital evidence management system becomes imperative for police researchers and investigators. This study aims at designing such a system using the design thinking method, which goes through five stages: empathy, definition, idea, prototype, and test. The result of the study is a web-based system prototype. The prototype user testing attains a system usability scale (SUS) value of 60. The SUS value means that the prototype is in the category of marginal low and indicates that the prototype does not meet the feasibility and needs improvement.","PeriodicalId":326094,"journal":{"name":"Khazanah Informatika : Jurnal Ilmu Komputer dan Informatika","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129103810","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-03DOI: 10.23917/khif.v8i1.15311
S. Suwarno
Since March 2020, due to the COVID-19 pandemic and in line with Merdeka Belajar - Kampus Merdeka, higher education institutions have conducted distance learning in asynchronous and synchronous modes, such as video meetings using Microsoft Teams and provide e-learning. In order to reach the goals and strategies of the higher education institutions, universities implement several control objectives within the COBIT 5 framework, so they can use and manage resources efficiently, provide the best education for students. This study aims to analyze the acceptance level of the COBIT implementation in higher education institutions by using the UTAUT model in E-Learning management, the use of Microsoft Teams and distance learning. This study uses a quantitative approach with a causal explanatory research design. Dissemination of the survey was conducted by simple random sampling at 6 (six) universities in Batam City. This study reveals that E-Learning management, the use of Microsoft Teams, and the application of distance learning together have a significant influence on the implementation of COBIT with an acceptance index of 85.5%, which refers to the satisfying category.
自2020年3月以来，由于2019冠状病毒病大流行，高等教育机构根据Merdeka Belajar - Kampus Merdeka，以异步和同步方式开展远程教育，例如使用Microsoft Teams进行视频会议，并提供电子学习。为了达到高等教育机构的目标和战略，大学在COBIT 5框架内实施了几个控制目标，这样他们就可以有效地使用和管理资源，为学生提供最好的教育。本研究旨在通过在E-Learning管理中使用UTAUT模型、使用Microsoft Teams和远程学习来分析高等教育机构对COBIT实施的接受程度。本研究采用定量方法和因果解释研究设计。在巴淡市的6所大学通过简单随机抽样进行了调查的传播。本研究发现，E-Learning管理、Microsoft Teams的使用和远程学习的应用共同对COBIT的实施产生了显著影响，接受指数为85.5%，属于满意类别。
{"title":"Application of the UTAUT Model for Acceptance Analysis of COBIT Implementation in E-Learning Management with Microsoft Teams on Distance Learning in Batam City","authors":"S. Suwarno","doi":"10.23917/khif.v8i1.15311","DOIUrl":"https://doi.org/10.23917/khif.v8i1.15311","url":null,"abstract":"Since March 2020, due to the COVID-19 pandemic and in line with Merdeka Belajar - Kampus Merdeka, higher education institutions have conducted distance learning in asynchronous and synchronous modes, such as video meetings using Microsoft Teams and provide e-learning. In order to reach the goals and strategies of the higher education institutions, universities implement several control objectives within the COBIT 5 framework, so they can use and manage resources efficiently, provide the best education for students. This study aims to analyze the acceptance level of the COBIT implementation in higher education institutions by using the UTAUT model in E-Learning management, the use of Microsoft Teams and distance learning. This study uses a quantitative approach with a causal explanatory research design. Dissemination of the survey was conducted by simple random sampling at 6 (six) universities in Batam City. This study reveals that E-Learning management, the use of Microsoft Teams, and the application of distance learning together have a significant influence on the implementation of COBIT with an acceptance index of 85.5%, which refers to the satisfying category.","PeriodicalId":326094,"journal":{"name":"Khazanah Informatika : Jurnal Ilmu Komputer dan Informatika","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131691363","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
- Blindness is a term to describe conditions of people who have visual impairments. When a visually impaired do an activity outside, he usually needs a stick to help them move. This study aims to develop a stick tracking that enable a family member to find the location of the blind when they are outside their home and can help the blind to travel. GPS (Global Positioning System) technology allows the stick to get a signal for its location coordinates. When a family member wants to get the location of the blind, he can send a text message with the keyword TRACKER to the mobile phone number of the stick. A GSM (Global System for Mobile Communication) module will send a reply containing the global coordinate, which Google Maps can visualize. In addition, the blind can actively send an emergency help signal to families if they have difficulty finding their way home. An emergency push button is available on the stick, which, if pressed, will send the coordinates to the family’s phone number in the form of a short text message. During travelling, blind people can identify obstacles in front of them thanks to an ultrasonic sensor system on the stick. The sensor can detect an object in the range of 100 cm. If the sensor detects an object less than 100 cm, a buzzer will emit an edible sound for the blind. Observations show that the developed stick works well with an average error on the GPS module at a level of 11.89 meters. It also shows a fluctuating percentage of ultrasonic sensor errors depending on the distance of objects.
{"title":"Blind People Stick Tracking Using Android Smartphone and GPS Technology","authors":"Rian Adi Chandra, Umi Fadlillah, Prasetyo Wibowo, Faizal Tegar Nanda Saputra, Reyhan Radditya Sulasyono","doi":"10.23917/khif.v8i1.15264","DOIUrl":"https://doi.org/10.23917/khif.v8i1.15264","url":null,"abstract":"- Blindness is a term to describe conditions of people who have visual impairments. When a visually impaired do an activity outside, he usually needs a stick to help them move. This study aims to develop a stick tracking that enable a family member to find the location of the blind when they are outside their home and can help the blind to travel. GPS (Global Positioning System) technology allows the stick to get a signal for its location coordinates. When a family member wants to get the location of the blind, he can send a text message with the keyword TRACKER to the mobile phone number of the stick. A GSM (Global System for Mobile Communication) module will send a reply containing the global coordinate, which Google Maps can visualize. In addition, the blind can actively send an emergency help signal to families if they have difficulty finding their way home. An emergency push button is available on the stick, which, if pressed, will send the coordinates to the family’s phone number in the form of a short text message. During travelling, blind people can identify obstacles in front of them thanks to an ultrasonic sensor system on the stick. The sensor can detect an object in the range of 100 cm. If the sensor detects an object less than 100 cm, a buzzer will emit an edible sound for the blind. Observations show that the developed stick works well with an average error on the GPS module at a level of 11.89 meters. It also shows a fluctuating percentage of ultrasonic sensor errors depending on the distance of objects.","PeriodicalId":326094,"journal":{"name":"Khazanah Informatika : Jurnal Ilmu Komputer dan Informatika","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128893605","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-03DOI: 10.23917/khif.v8i1.15531
Ari Peryanto, A. Yudhana, R. Umar
- Flowers are among the raw materials in many industries including the pharmaceuticals and cosmetics. Manual classification of flowers requires expert judgment of a botanist and can be time consuming and inconsistent. The ability to classify flowers using computers and technology is the right solution to solve this problem. There are two algorithms that are popular in image classification, namely Convolutional Neural Network (CNN) and Support Vector Machine (SVM). CNN is one of deep neural network classification algorithms while SVM is one of machine learning algorithm. This research was an effort to determine the best performer of the two methods in flower image classification. Our observation suggests that CNN outperform SVM in flower image classification. CNN gives an accuracy of 91.6%, precision of 91.6%, recall of 91.6% and F1 Score of 91.6%.
{"title":"Convolutional Neural Network and Support Vector Machine in Classification of Flower Images","authors":"Ari Peryanto, A. Yudhana, R. Umar","doi":"10.23917/khif.v8i1.15531","DOIUrl":"https://doi.org/10.23917/khif.v8i1.15531","url":null,"abstract":"- Flowers are among the raw materials in many industries including the pharmaceuticals and cosmetics. Manual classification of flowers requires expert judgment of a botanist and can be time consuming and inconsistent. The ability to classify flowers using computers and technology is the right solution to solve this problem. There are two algorithms that are popular in image classification, namely Convolutional Neural Network (CNN) and Support Vector Machine (SVM). CNN is one of deep neural network classification algorithms while SVM is one of machine learning algorithm. This research was an effort to determine the best performer of the two methods in flower image classification. Our observation suggests that CNN outperform SVM in flower image classification. CNN gives an accuracy of 91.6%, precision of 91.6%, recall of 91.6% and F1 Score of 91.6%.","PeriodicalId":326094,"journal":{"name":"Khazanah Informatika : Jurnal Ilmu Komputer dan Informatika","volume":"68 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115776846","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-03-03DOI: 10.23917/khif.v8i1.15533
Lukman Reza, S. Sunardi, H. Herman
Implementation of Academic Information System (AIS) at the Radya Binatama Communication Academy (AKRB) had some problems for some users. That problems are known from interviews with several AIS users. This caused a delay in data exchange with other divisions. This study aims to assess the implementation of AIS to the AKRB by using the Unified Theory of Acceptance and Use of Technology (UTAUT) method. UTAUT has four main constructs that affect user acceptance namely performance expectations, effort expectations, social influences, and facilitating conditions. The data were obtained from distributing questionnaires to all AIS users as many as 40 respondents. Then the data is processed using Structural Equation Modeling (SEM) techniques with the help of SmartPLS. The results of the analysis show that only construct facilitating conditions is valid with a t-statistic value of 2.733. While the other three constructs have values that are in the range of invalid values between -1.96 to 1.96 with the values of each construct being 1.891, 0.050, 1.440. It can be concluded that the application of SIA in AKRB has not been well received by all AIS users. Therefore, it is necessary to conduct an evaluation that represents the other three constructs.
{"title":"Academic Information System Assessment of AKRB Yogyakarta Using UTAUT","authors":"Lukman Reza, S. Sunardi, H. Herman","doi":"10.23917/khif.v8i1.15533","DOIUrl":"https://doi.org/10.23917/khif.v8i1.15533","url":null,"abstract":"Implementation of Academic Information System (AIS) at the Radya Binatama Communication Academy (AKRB) had some problems for some users. That problems are known from interviews with several AIS users. This caused a delay in data exchange with other divisions. This study aims to assess the implementation of AIS to the AKRB by using the Unified Theory of Acceptance and Use of Technology (UTAUT) method. UTAUT has four main constructs that affect user acceptance namely performance expectations, effort expectations, social influences, and facilitating conditions. The data were obtained from distributing questionnaires to all AIS users as many as 40 respondents. Then the data is processed using Structural Equation Modeling (SEM) techniques with the help of SmartPLS. The results of the analysis show that only construct facilitating conditions is valid with a t-statistic value of 2.733. While the other three constructs have values that are in the range of invalid values between -1.96 to 1.96 with the values of each construct being 1.891, 0.050, 1.440. It can be concluded that the application of SIA in AKRB has not been well received by all AIS users. Therefore, it is necessary to conduct an evaluation that represents the other three constructs.","PeriodicalId":326094,"journal":{"name":"Khazanah Informatika : Jurnal Ilmu Komputer dan Informatika","volume":"28 2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126831529","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-10-30DOI: 10.23917/khif.v6i2.11058
Fitri Purwaningtias, Maria Ulfa, Febi Franata
Junjung Biru Waste Bank conducts a selection of the best member biennially. The process is crucial, but it does not have a supporting system, which poses problems emerging from data redundancies and data loss. Among the problem is the difficulty for administrators in summarizing data of members who have transactions. To solve the problem, we devised and implemented a decision support system using the CPI (Composite Performance Index) method. The criteria are the amount of balance and active saving during a six-month interval. The results of this research is a web-based decision support system that produces a ranking order of members, which helps in selecting the best member.
{"title":"Decision Support System for Selection of the Best Member at Junjung Biru Waste Bank Using the Composite Performance Index (CPI)","authors":"Fitri Purwaningtias, Maria Ulfa, Febi Franata","doi":"10.23917/khif.v6i2.11058","DOIUrl":"https://doi.org/10.23917/khif.v6i2.11058","url":null,"abstract":"Junjung Biru Waste Bank conducts a selection of the best member biennially. The process is crucial, but it does not have a supporting system, which poses problems emerging from data redundancies and data loss. Among the problem is the difficulty for administrators in summarizing data of members who have transactions. To solve the problem, we devised and implemented a decision support system using the CPI (Composite Performance Index) method. The criteria are the amount of balance and active saving during a six-month interval. The results of this research is a web-based decision support system that produces a ranking order of members, which helps in selecting the best member.","PeriodicalId":326094,"journal":{"name":"Khazanah Informatika : Jurnal Ilmu Komputer dan Informatika","volume":"68 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128777553","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-10-30DOI: 10.23917/khif.v6i2.11013
B. Sumanto, M. Fakhrurrifqi
Fish meat is a source of minerals and protein and contains excellent nutrients for the human body. However, non-fresh (rotting) fish are sometimes in the market for sale. Consuming rotting fish puts people at risk of getting diseases. This paper describes research to build a smelling device (e-nose) to identify fish freshness. It aims at detecting unsafe fish flesh to sort them out from being sold. We cut red snapper into cubes and put them into an open space at room temperature for five days. During the period, a gas sensor array acquired data of gas smell from the rotting fish. The output voltage of the sensors was processed using the differential baseline method. Later, feature extraction took the maximum value from the response of the gas sensor array, while the Principle Component Analysis (PCA) method identified the pattern. The results suggest that the gas sensor array responds to changes in the smell of fish meat that undergo a decay process. The PCA method is capable of recognizing the pattern of the maximum value characteristic of the gas sensor array response, as evidenced by the cumulative values of PC1 and PC2 reaching 95.95% with an accuracy rate of 98.2%. It shows the correlation between the aroma profiles of fish meat during the spoilage process, which produces a sharper aroma due to microbiological growth in the fish meat.
{"title":"Utilization of Gas Sensor Array and Principal Component Analysis to Identify Fish Decomposition Level","authors":"B. Sumanto, M. Fakhrurrifqi","doi":"10.23917/khif.v6i2.11013","DOIUrl":"https://doi.org/10.23917/khif.v6i2.11013","url":null,"abstract":"Fish meat is a source of minerals and protein and contains excellent nutrients for the human body. However, non-fresh (rotting) fish are sometimes in the market for sale. Consuming rotting fish puts people at risk of getting diseases. This paper describes research to build a smelling device (e-nose) to identify fish freshness. It aims at detecting unsafe fish flesh to sort them out from being sold. We cut red snapper into cubes and put them into an open space at room temperature for five days. During the period, a gas sensor array acquired data of gas smell from the rotting fish. The output voltage of the sensors was processed using the differential baseline method. Later, feature extraction took the maximum value from the response of the gas sensor array, while the Principle Component Analysis (PCA) method identified the pattern. The results suggest that the gas sensor array responds to changes in the smell of fish meat that undergo a decay process. The PCA method is capable of recognizing the pattern of the maximum value characteristic of the gas sensor array response, as evidenced by the cumulative values of PC1 and PC2 reaching 95.95% with an accuracy rate of 98.2%. It shows the correlation between the aroma profiles of fish meat during the spoilage process, which produces a sharper aroma due to microbiological growth in the fish meat.","PeriodicalId":326094,"journal":{"name":"Khazanah Informatika : Jurnal Ilmu Komputer dan Informatika","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121997845","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-10-30DOI: 10.23917/khif.v6i2.11344
F. A. Muqtadiroh, Anisah Herdiyanti, N. Puspitasari
This paper aims to understand the behavioral intentions of students in using e-learning in a public university in Indonesia. We apply the e-learning quality model to observe the quality factors that trigger intentions. The quality factors include assurance, empathy, responsiveness, reliability, and website content. Understanding how these quality learning factors may affect a student’s behavior intention to use e-learning is important to bring e-learning implementation success. We collected 502 responses from university students at a public university in Indonesia that implements a Moodle-based e-learning platform – namely ShareITS. Out of 5 (five) quality learning factors, we found only 2 (two) that significantly affect the e-learning quality – i.e., responsiveness and website content. The result shows that the teacher-student engagement in the e-learning platform and also the better visual design of ShareITS can improve the quality of the e-learning platform.
{"title":"The e-Learning Quality Model to Examine Students’ Behavioral Intention to Use Online Learning Platform in a Higher Education Institution","authors":"F. A. Muqtadiroh, Anisah Herdiyanti, N. Puspitasari","doi":"10.23917/khif.v6i2.11344","DOIUrl":"https://doi.org/10.23917/khif.v6i2.11344","url":null,"abstract":"This paper aims to understand the behavioral intentions of students in using e-learning in a public university in Indonesia. We apply the e-learning quality model to observe the quality factors that trigger intentions. The quality factors include assurance, empathy, responsiveness, reliability, and website content. Understanding how these quality learning factors may affect a student’s behavior intention to use e-learning is important to bring e-learning implementation success. We collected 502 responses from university students at a public university in Indonesia that implements a Moodle-based e-learning platform – namely ShareITS. Out of 5 (five) quality learning factors, we found only 2 (two) that significantly affect the e-learning quality – i.e., responsiveness and website content. The result shows that the teacher-student engagement in the e-learning platform and also the better visual design of ShareITS can improve the quality of the e-learning platform.","PeriodicalId":326094,"journal":{"name":"Khazanah Informatika : Jurnal Ilmu Komputer dan Informatika","volume":"322 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115918988","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-10-27DOI: 10.23917/khif.v6i2.10520
I. Waspada, N. Bahtiar, P. W. Wirawan, Bagus Dwi Ari Awan
Losses incurred due to fraud on e-commerce transactions, especially those based on credit cards, continue to increase, resulting in large losses each year. One mechanism to minimize the risk of fraudulent credit card transactions is to utilize a detection technique for ongoing transactions. Credit card transaction data in its original state does not have a label, and the amount of fraud data on the training data is very small so that it belongs to a very unbalanced category, and the pattern of fraud continues to change. Isolation forest is an unsupervised algorithm that is efficient in detecting anomalies. Several techniques can be applied to improve the performance of the Isolation forest model. Previous studies used the ROC-AUC metric in analyzing the performance of Isolation Forests, which could provide incorrect information. This study made two contributions; the first is to present a performance analysis with both the ROC-AUC and AUCPR. Thus, it can be seen that the high ROC-AUC value does not guarantee the model has the reliability in detecting fraud. In comparison, the information provided through AUCPR is more appropriate to describe the ability of the model to capture data fraud. The second contribution is to propose several techniques that can be applied to improve the performance of the Isolation forest model, namely to optimize the determination of the amount of training data, feature selection, the amount of fraud contamination, and setting hyper-parameters in the modeling stage (training). Experiments were carried out using a real-life dataset from ULB. The best results are obtained when the validation data split ratio is 60:40, using the five most important features, using only 60% of fraud data, and setting hyper-parameters with the number of trees 100, 128 sample maximum, and 0.001 contamination. The validation performance of this model is precision 0.809917, recall 0.710145, F1-score 0.756757, ROC-AUC 0.969728, and AUCPR 0.637993, while for Testing results obtained precision 0.807143, recall 0.763514, F1-score 0.784722, ROC-AUC 0.97371, and AUCPR 0.759228.
{"title":"Performance Analysis of Isolation Forest Algorithm in Fraud Detection of Credit Card Transactions","authors":"I. Waspada, N. Bahtiar, P. W. Wirawan, Bagus Dwi Ari Awan","doi":"10.23917/khif.v6i2.10520","DOIUrl":"https://doi.org/10.23917/khif.v6i2.10520","url":null,"abstract":"Losses incurred due to fraud on e-commerce transactions, especially those based on credit cards, continue to increase, resulting in large losses each year. One mechanism to minimize the risk of fraudulent credit card transactions is to utilize a detection technique for ongoing transactions. Credit card transaction data in its original state does not have a label, and the amount of fraud data on the training data is very small so that it belongs to a very unbalanced category, and the pattern of fraud continues to change. Isolation forest is an unsupervised algorithm that is efficient in detecting anomalies. Several techniques can be applied to improve the performance of the Isolation forest model. Previous studies used the ROC-AUC metric in analyzing the performance of Isolation Forests, which could provide incorrect information. This study made two contributions; the first is to present a performance analysis with both the ROC-AUC and AUCPR. Thus, it can be seen that the high ROC-AUC value does not guarantee the model has the reliability in detecting fraud. In comparison, the information provided through AUCPR is more appropriate to describe the ability of the model to capture data fraud. The second contribution is to propose several techniques that can be applied to improve the performance of the Isolation forest model, namely to optimize the determination of the amount of training data, feature selection, the amount of fraud contamination, and setting hyper-parameters in the modeling stage (training). Experiments were carried out using a real-life dataset from ULB. The best results are obtained when the validation data split ratio is 60:40, using the five most important features, using only 60% of fraud data, and setting hyper-parameters with the number of trees 100, 128 sample maximum, and 0.001 contamination. The validation performance of this model is precision 0.809917, recall 0.710145, F1-score 0.756757, ROC-AUC 0.969728, and AUCPR 0.637993, while for Testing results obtained precision 0.807143, recall 0.763514, F1-score 0.784722, ROC-AUC 0.97371, and AUCPR 0.759228.","PeriodicalId":326094,"journal":{"name":"Khazanah Informatika : Jurnal Ilmu Komputer dan Informatika","volume":"46 1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130768746","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-10-27DOI: 10.23917/khif.v6i2.10867
D. Susanto, A. Fadlil, A. Yudhana
Goats are livestock that is financially very attractive to rural Indonesian. Efforts to solve problems related to goat farming are necessary. One of them is maintaining the health of the cattle by knowing how to cope with disease-stricken goats. Goat productivity will decrease if the treatment of the disease is sub-optimal. Goat diseases are very diverse, ranging from mild to severe. Breeders themselves can traditionally treat several diseases without the involvement of veterinarians or experts. However, a larger number of diseases need treatment with the help of experts. Expert systems are a potential solution to help farmers. It will automatically suggest decisions or conclusions in solving a problem. This study observes an expert system built using the Certainty Factor combined with Forward-Chaining. By combining the two methods, the information generated may discover the type of disease and suggest its management effectively with a high degree of certainty. The system can expectedly become a reference for goat breeders to consult about their goat livestock diseases. The knowledge base of the system uses 21 types of symptoms, eight types of diseases, and their solutions. The user does not need to input the belief value and the disbelief value that is usually input in the expert system. By involving the admin as a knowledge base processor, the correctness of the conveyed information maintains.
{"title":"Application of the Certainty Factor and Forward Chaining Methods to a Goat Disease Expert System","authors":"D. Susanto, A. Fadlil, A. Yudhana","doi":"10.23917/khif.v6i2.10867","DOIUrl":"https://doi.org/10.23917/khif.v6i2.10867","url":null,"abstract":"Goats are livestock that is financially very attractive to rural Indonesian. Efforts to solve problems related to goat farming are necessary. One of them is maintaining the health of the cattle by knowing how to cope with disease-stricken goats. Goat productivity will decrease if the treatment of the disease is sub-optimal. Goat diseases are very diverse, ranging from mild to severe. Breeders themselves can traditionally treat several diseases without the involvement of veterinarians or experts. However, a larger number of diseases need treatment with the help of experts. Expert systems are a potential solution to help farmers. It will automatically suggest decisions or conclusions in solving a problem. This study observes an expert system built using the Certainty Factor combined with Forward-Chaining. By combining the two methods, the information generated may discover the type of disease and suggest its management effectively with a high degree of certainty. The system can expectedly become a reference for goat breeders to consult about their goat livestock diseases. The knowledge base of the system uses 21 types of symptoms, eight types of diseases, and their solutions. The user does not need to input the belief value and the disbelief value that is usually input in the expert system. By involving the admin as a knowledge base processor, the correctness of the conveyed information maintains.","PeriodicalId":326094,"journal":{"name":"Khazanah Informatika : Jurnal Ilmu Komputer dan Informatika","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115005865","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：