Luh Kesuma Wardhani
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Luh Kesuma Wardhani has a total of 23 co-authors. Here are the top co-authors by number of shared publications:
- Husni Teja Sukmana
Informatics Department, Faculty of Science and Engineering, Universitas Islam Negeri Syarif Hidayatullah, Jakarta, Indonesia
3 shared publications Elin Haerani
Departement of Informatics, Faculty of Science and Technology, UIN Sultan Syarif Kasim Riau
2 shared publicationsNashrul Hakiem
Department of Informatics Faculty of Science and Technology UIN Syarif Hidayatullah Jakarta
2 shared publicationsAmanda Ayu Larasati
1 shared publicationAngga Kurniawan
Universitas Islam Negeri Syarif Hidayatullah Jakarta
1 shared publicationAnif Hanifa Setyaningrum
1 shared publication
Latest Publications
- 31 December 2025
- Published by Universitas Suryakancana in Media Jurnal Informatika
- Vol. 17 (2) , 367-378
Multi-Class Fault Detection under Class-Imbalance in Wireless Sensor Network Using Random Undersampling and Extra Trees
Wireless Sensor Networks (WSNs) are widely used in various monitoring applications, including environmental observation, smart infrastructure, and Internet of Things (IoT) systems. Despite their widespread adoption, WSNs are highly susceptible to data errors caused by sensor degradation, hardware malfunctions, environmental disturbances, and communication issues. These faults can significantly reduce data reliability and lead to incorrect system decisions if not properly handled. This study proposes a multi-class data-fault detection approach for WSNs under imbalanced data conditions by integrating Random Undersampling (RUS) with the Extra-Trees classification algorithm. The proposed framework aims to address the class imbalance problem commonly found in sensor fault datasets while improving fault detection performance across multiple fault types. Experiments were conducted using a WSN dataset containing temperature and humidity measurements, in which three fault types: Bias, Drift, and Spike were analyzed alongside normal sensor data. The experimental results demonstrate that Random Undersampling leads to a substantial improvement in classification performance. Without RUS, the Extra-Trees classifier achieved an accuracy of 48% and failed to detect spike faults. After applying RUS, classification accuracy increased to 91%, accompanied by balanced precision, recall, and F1-score values across all classes. These findings indicate that the combination of Random Undersampling and Extra-Trees provides an effective and reliable solution for multi-class data fault detection in WSN environments.
- 31 December 2025
- Published by Universitas Suryakancana in Media Jurnal Informatika
- Vol. 17 (2) , 367-378
Multi-Class Fault Detection under Class-Imbalance in Wireless Sensor Network Using Random Undersampling and Extra Trees
Wireless Sensor Networks (WSNs) are widely used in various monitoring applications, including environmental observation, smart infrastructure, and Internet of Things (IoT) systems. Despite their widespread adoption, WSNs are highly susceptible to data errors caused by sensor degradation, hardware malfunctions, environmental disturbances, and communication issues. These faults can significantly reduce data reliability and lead to incorrect system decisions if not properly handled. This study proposes a multi-class data-fault detection approach for WSNs under imbalanced data conditions by integrating Random Undersampling (RUS) with the Extra-Trees classification algorithm. The proposed framework aims to address the class imbalance problem commonly found in sensor fault datasets while improving fault detection performance across multiple fault types. Experiments were conducted using a WSN dataset containing temperature and humidity measurements, in which three fault types: Bias, Drift, and Spike were analyzed alongside normal sensor data. The experimental results demonstrate that Random Undersampling leads to a substantial improvement in classification performance. Without RUS, the Extra-Trees classifier achieved an accuracy of 48% and failed to detect spike faults. After applying RUS, classification accuracy increased to 91%, accompanied by balanced precision, recall, and F1-score values across all classes. These findings indicate that the combination of Random Undersampling and Extra-Trees provides an effective and reliable solution for multi-class data fault detection in WSN environments.
- 1 September 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 38, 74-78
Performance Analysis of Naïve Bayes Classifier Algorithm with Chi-Square and Confix Stripping Stemmer Selection Features In Hadits Translation Classification System
Hadith is the second source of Islamic law after the Qur'an. Hadith classification supports the authentication of a hadith. The hadith classification process is carried out by the computer research community, while hadith authentication is carried out by hadith scholars. Text classification is grouping a certain text into one of the classes given a category based on the content of the text. One of the text mining classification algorithms is Naïve Bayes Classifier. In this study, researchers simulated a combination of the Naïve Bayes Classifier algorithm with Confix Stripping Stemmer for stemming, and Chi-Square as a feature selection algorithm to classify the translation of Sahih Bukhari's hadith into three classes, namely suggestions, prohibitions, and information with 300 data training data, and 30 testing data. System performance testing uses a confusion matrix by calculating the value of accuracy, precision, recall, and f-measure. The results show the best performance is a combination of Naïve Bayes Classifier and Confix Stripping Stemmer.
- 1 September 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 38, 74-78
Performance Analysis of Naïve Bayes Classifier Algorithm with Chi-Square and Confix Stripping Stemmer Selection Features In Hadits Translation Classification System
Hadith is the second source of Islamic law after the Qur'an. Hadith classification supports the authentication of a hadith. The hadith classification process is carried out by the computer research community, while hadith authentication is carried out by hadith scholars. Text classification is grouping a certain text into one of the classes given a category based on the content of the text. One of the text mining classification algorithms is Naïve Bayes Classifier. In this study, researchers simulated a combination of the Naïve Bayes Classifier algorithm with Confix Stripping Stemmer for stemming, and Chi-Square as a feature selection algorithm to classify the translation of Sahih Bukhari's hadith into three classes, namely suggestions, prohibitions, and information with 300 data training data, and 30 testing data. System performance testing uses a confusion matrix by calculating the value of accuracy, precision, recall, and f-measure. The results show the best performance is a combination of Naïve Bayes Classifier and Confix Stripping Stemmer.
- Open Access
- 13 April 2022
- Published by Faculty of Computer Science, Universitas Indonesia in Jurnal Sistem Informasi
- Vol. 18 (1) , 31-44
Whistleblowing System Deployment using ITIL v3 Framework
The Indonesian government has been urged to implement Whistleblowing systems for public services in order to achieve integrity zone development. Almost all service desk software provides user convenience as well as manages the service process in an organisation using standards based on the Information Technology Infrastructure Library (ITIL) v3. The service operation of Whistleblowing systems and the service desk systems are almost the same. However, currently only a few studies have examined the implementation of Whistleblowing systems which uses such a framework in a public higher education institution. The aim of this study is to design and develop a Whistleblowing system which is expected to fulfil the IT Service Management (ITSM) standard and to perform an integration process amongst the community and students. The software was designed and tested in the Universitas Islam Negeri Syarif Hidayatullah Jakarta using the ITIL v3 standards. The result showed that the efficiency parameter value produced a page speed point of 91 % (A). The minimum score was 91 % (A) and the fully loaded time was 2.4 seconds. A portability test showed a value of 100 % for the mobile-friendly score while the reliability test with a stress test also showed at 100 %. The alpha testing result was a very good predicate while beta testing received an acceptable score, good category, and classified as grade C. In conclusion, based on the abovementioned tests, this shows that the system can be used to manage student services, lecturers and staff, as well as to integrate the current service within the community.
- 13 April 2022
- Published by Faculty of Computer Science, Universitas Indonesia in Jurnal Sistem Informasi
- Vol. 18 (1) , 31-44
Open AccessWhistleblowing System Deployment using ITIL v3 Framework
The Indonesian government has been urged to implement Whistleblowing systems for public services in order to achieve integrity zone development. Almost all service desk software provides user convenience as well as manages the service process in an organisation using standards based on the Information Technology Infrastructure Library (ITIL) v3. The service operation of Whistleblowing systems and the service desk systems are almost the same. However, currently only a few studies have examined the implementation of Whistleblowing systems which uses such a framework in a public higher education institution. The aim of this study is to design and develop a Whistleblowing system which is expected to fulfil the IT Service Management (ITSM) standard and to perform an integration process amongst the community and students. The software was designed and tested in the Universitas Islam Negeri Syarif Hidayatullah Jakarta using the ITIL v3 standards. The result showed that the efficiency parameter value produced a page speed point of 91 % (A). The minimum score was 91 % (A) and the fully loaded time was 2.4 seconds. A portability test showed a value of 100 % for the mobile-friendly score while the reliability test with a stress test also showed at 100 %. The alpha testing result was a very good predicate while beta testing received an acceptable score, good category, and classified as grade C. In conclusion, based on the abovementioned tests, this shows that the system can be used to manage student services, lecturers and staff, as well as to integrate the current service within the community.
- Open Access
- 1 December 2018
- Published by Universitas Ahmad Dahlan in TELKOMNIKA (Telecommunication Computing Electronics and Control)
- Vol. 16 (6) , 2733-2739
Speech Recognition Application for the Speech Impaired using the Android-based Google Cloud Speech API
Speech Recognition Application for the Speech Impaired using the Android-based Google Cloud Speech API
- 1 December 2018
- Published by Universitas Ahmad Dahlan in TELKOMNIKA (Telecommunication Computing Electronics and Control)
- Vol. 16 (6) , 2733-2739
Open AccessSpeech Recognition Application for the Speech Impaired using the Android-based Google Cloud Speech API
Speech Recognition Application for the Speech Impaired using the Android-based Google Cloud Speech API
- 1 August 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Optimization of multiple depot vehicle routing problem (MDVRP) on perishable product distribution by using genetic algorithm and fuzzy logic controller (FLC)
Distribution has a large portion of the total cost of sales of goods or product. It cause high cost transport and risk to the quality of the goods which are distributed, especially perishable product. Perishable product quality is greatly influenced by distribution, thus when the distribution not optimal, it will risk very big deal of broken product. In connection with this matter, the distribution process needed to optimize especially to perishable product. Optimal distribution problem can be solved using Multiple Depot Vehicle Distribution Routing Problem (MDVRP) model, which was development of Vehicle Routing Problem (VRP). To get the optimal solution of MDVRP, genetic Algorithm (GA) was used. Using GA, more than one alternative solution with good fitness value were gained. Then Fuzzy Logic Controller (FLC) was used to control these parameter values in order not to trap into local optimum. After new pm and pc value were obtained, the GA process was conducted to get the shortest route. The result of this research showed that the value of the probability of crossover and mutation affected the fitness values. The higher fitness value, the smaller the distance which reached. It will reduce the risk of rot caused by travel time. Testing results showed some alternative solutions which is not much different in fitness value. Level of optimality increased when FLC process were conducted. From 10 times evaluation reached 89, 9%.
- 1 August 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Optimization of multiple depot vehicle routing problem (MDVRP) on perishable product distribution by using genetic algorithm and fuzzy logic controller (FLC)
Distribution has a large portion of the total cost of sales of goods or product. It cause high cost transport and risk to the quality of the goods which are distributed, especially perishable product. Perishable product quality is greatly influenced by distribution, thus when the distribution not optimal, it will risk very big deal of broken product. In connection with this matter, the distribution process needed to optimize especially to perishable product. Optimal distribution problem can be solved using Multiple Depot Vehicle Distribution Routing Problem (MDVRP) model, which was development of Vehicle Routing Problem (VRP). To get the optimal solution of MDVRP, genetic Algorithm (GA) was used. Using GA, more than one alternative solution with good fitness value were gained. Then Fuzzy Logic Controller (FLC) was used to control these parameter values in order not to trap into local optimum. After new pm and pc value were obtained, the GA process was conducted to get the shortest route. The result of this research showed that the value of the probability of crossover and mutation affected the fitness values. The higher fitness value, the smaller the distance which reached. It will reduce the risk of rot caused by travel time. Testing results showed some alternative solutions which is not much different in fitness value. Level of optimality increased when FLC process were conducted. From 10 times evaluation reached 89, 9%.
- 31 July 2017
- Published by NADIA in International Journal of Control and Automation
- Vol. 10 (7) , 203-216
The Evaluation of ITSM Open Source Software for Small Medium Organizations Based on ITIL v.3 Criteria using AHP Method
- 31 July 2017
- Published by NADIA in International Journal of Control and Automation
- Vol. 10 (7) , 203-216
The Evaluation of ITSM Open Source Software for Small Medium Organizations Based on ITIL v.3 Criteria using AHP Method
- 1 November 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 160-165
Development Decision Support System of Choosing Medicine Using TOPSIS Method (Case Study: RSIA Tiara)
Nowadays, choosing medicine is one of the encountered problem in daily life. In RSIA Tiara, after patient gets checked, doctor will diagnose what is the patient's illness and doctor will give them medicine prescription. Patient then needs to redeem the medicines at pharmacy. The problem arises when there are patients who do not want to redeem presciptions. The reason why patients do not want to redeem the prescriptions is because of the criteria of the medicines are not suitable to what patients want. The criteria of medicines are price, type, quality, taste, and comfort. Therefore, to overcome this problem, a decision support system is proposed. The method the author is used in this research is TOPSIS method.
- 1 November 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 160-165
Development Decision Support System of Choosing Medicine Using TOPSIS Method (Case Study: RSIA Tiara)
Nowadays, choosing medicine is one of the encountered problem in daily life. In RSIA Tiara, after patient gets checked, doctor will diagnose what is the patient's illness and doctor will give them medicine prescription. Patient then needs to redeem the medicines at pharmacy. The problem arises when there are patients who do not want to redeem presciptions. The reason why patients do not want to redeem the prescriptions is because of the criteria of the medicines are not suitable to what patients want. The criteria of medicines are price, type, quality, taste, and comfort. Therefore, to overcome this problem, a decision support system is proposed. The method the author is used in this research is TOPSIS method.
- 29 September 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-4
Application of unsupervised K Nearest Neighbor (UNN) and Learning Vector Quantization (LVQ) methods in predicting Rupiah to Dollar
One of the factors in a country's economy is the exchange value of the currency towards another currency. The exchange value of Rupiah towards Dollar (USA) can quickly change depending on the environmental conditions and has a huge impact for the Indonesian Government. In this research, Learning Vector Quantization (LVQ) and Unsupervised K Nearest Network (UNN) was implemented in predicting the currency value towards dollar. The UNN method was used to predict the selling value of the currency, the LVQ method was used to predict the buying value of the currency. The input data that is used is the selling, buying and interest data times series of the currency from the central back of the United States. From the research result and discussions that was made, UNN can achieve the lowest MAPE, which is 1,544% with the amount of data as much as 25 and the LVQ algorithm can accurately achieve a forecast with the amount of data as much as 25 with the learning rate of 0,075. The amount of trained data and the many patterns that exist in one LVQ class method can affect the result of the study and the result of the system.
- 29 September 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-4
Application of unsupervised K Nearest Neighbor (UNN) and Learning Vector Quantization (LVQ) methods in predicting Rupiah to Dollar
One of the factors in a country's economy is the exchange value of the currency towards another currency. The exchange value of Rupiah towards Dollar (USA) can quickly change depending on the environmental conditions and has a huge impact for the Indonesian Government. In this research, Learning Vector Quantization (LVQ) and Unsupervised K Nearest Network (UNN) was implemented in predicting the currency value towards dollar. The UNN method was used to predict the selling value of the currency, the LVQ method was used to predict the buying value of the currency. The input data that is used is the selling, buying and interest data times series of the currency from the central back of the United States. From the research result and discussions that was made, UNN can achieve the lowest MAPE, which is 1,544% with the amount of data as much as 25 and the LVQ algorithm can accurately achieve a forecast with the amount of data as much as 25 with the learning rate of 0,075. The amount of trained data and the many patterns that exist in one LVQ class method can affect the result of the study and the result of the system.
- 29 September 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Application of genetic algorithm for class scheduling (Case study: Faculty of science and technology UIN Jakarta)
The objective of this research is about building a class scheduling application using genetic algorithm. The parameters used in genetic algorithms namely: iteration, PM (Probability Mutation), PC (Probability crossover). In Faculty of Science and Technology UIN Jakarta, class schedule built by a staff. The staff faces numbers of obstacles when building the schedule, among others: the limitations of the classroom and schedule of lectures, and large number of students. However, it is able to overcome the obstacles but it takes quite a long time to build. To optimize the tasks, an application using genetic algorithm is suggested. In this research, interviews and literature studies are used as research methodology, Waterfall as application development methods and genetic algorithms for scheduling. This application was developed using HTML, PHP, and MySQL. To get the optimal parameter values,_some variance of iteration (ns), PC, PM values were tested. The results show that the optimal value for gaining the best schedule are maximum of 20 iterations, PC 0.8 and PM 0.01. Using this values, the class scheduling application generates approximately 1,201 without any clash among the data and this application can facilitate the building of class schedule.
- 29 September 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Application of genetic algorithm for class scheduling (Case study: Faculty of science and technology UIN Jakarta)
The objective of this research is about building a class scheduling application using genetic algorithm. The parameters used in genetic algorithms namely: iteration, PM (Probability Mutation), PC (Probability crossover). In Faculty of Science and Technology UIN Jakarta, class schedule built by a staff. The staff faces numbers of obstacles when building the schedule, among others: the limitations of the classroom and schedule of lectures, and large number of students. However, it is able to overcome the obstacles but it takes quite a long time to build. To optimize the tasks, an application using genetic algorithm is suggested. In this research, interviews and literature studies are used as research methodology, Waterfall as application development methods and genetic algorithms for scheduling. This application was developed using HTML, PHP, and MySQL. To get the optimal parameter values,_some variance of iteration (ns), PC, PM values were tested. The results show that the optimal value for gaining the best schedule are maximum of 20 iterations, PC 0.8 and PM 0.01. Using this values, the class scheduling application generates approximately 1,201 without any clash among the data and this application can facilitate the building of class schedule.
- 29 September 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-7
Android application for diagnosing general symptoms of disease using forward chaining method
Sometimes people feel doubtful about things that happened to their physical health. Especially when there is no place to ask about the pain they feel or experiencing. So usually and are more often they ignoring and underestimate the symptoms they are experiencing. Generally someone knows that minor illnesses such as colds or diarrhea, do not require special examination or complex treatment. The problem is, minor illness could be an indicator of a serious disease. Therefore one must know the symptoms of disease they feel with diagnose the symptoms. To facilitate someone to diagnose the symptoms they experienced, we develop an Android based application that can provide solutions for what they should do when they are experiencing these symptoms, whether they could do their own treatment at home, or should be immediately checked by a doctor. This application using inference forward chaining method. By using this application, diagnostic results of disease can be detected through the consultation process or answering the questions provided by the system quickly and effectively with an Android smartphone as the media.
- 29 September 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-7
Android application for diagnosing general symptoms of disease using forward chaining method
Sometimes people feel doubtful about things that happened to their physical health. Especially when there is no place to ask about the pain they feel or experiencing. So usually and are more often they ignoring and underestimate the symptoms they are experiencing. Generally someone knows that minor illnesses such as colds or diarrhea, do not require special examination or complex treatment. The problem is, minor illness could be an indicator of a serious disease. Therefore one must know the symptoms of disease they feel with diagnose the symptoms. To facilitate someone to diagnose the symptoms they experienced, we develop an Android based application that can provide solutions for what they should do when they are experiencing these symptoms, whether they could do their own treatment at home, or should be immediately checked by a doctor. This application using inference forward chaining method. By using this application, diagnostic results of disease can be detected through the consultation process or answering the questions provided by the system quickly and effectively with an Android smartphone as the media.