Articles
98 Documents
<![CDATA[Performance Analysis of the Hybrid Voting Method on the Classification of the Number of Cases of Dengue Fever ]]>
<![CDATA[arief rahman]]>;
<![CDATA[Sri Suryani Prasetiyowati]]>
<![CDATA[ International Journal on Information and Communication Technology (IJoICT) Vol. 8 No. 1 (2022): June 2022 ]]>
Publisher : <![CDATA[School of Computing, Telkom University ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.21108/ijoict.v8i1.614
<![CDATA[Dengue hemorrhagic fever (DHF) is a health problem in Indonesia. The region in Indonesia that has the highest number of cases in West Java with the highest ranking with 10,772 cases. The city of Bandung is recorded to have the highest number of cases at this time, namely 4,424 cases. Dengue fever can be caused by high rainfall. Judging from the high number of cases and fluctuations that occur, it is necessary to predict the spread of the disease so that in the future it can be anticipated by the government. Prediction of the spread of dengue fever in the city of Bandung using various classification algorithms has been done. Therefore, the author wants to make a new breakthrough by using hybrid ensemble learning using a hard voting method from three classification methods, namely Support Vector Machine (SVM), K-Nearest Neighbor (KNN), and Decision Tree (DT). Using the Bandung City DHF disease dataset from 2012 to 2018. The results obtained using the Support Vector Machine (SVM), K-Nearest Neighbor (KNN) and Decision Tree (DT) were 84%, 87%, 79%. to improve the classification accuracy of the three methods using a hybrid classification with the hard voting method to get 91% results.]]>
<![CDATA[Overcoming Data Imbalance Problems in Sexual Harassment Classification with SMOTE ]]>
<![CDATA[Aji Gautama Putrada]]>;
<![CDATA[Irfan Dwi Wijaya]]>;
<![CDATA[Dita Oktaria]]>
<![CDATA[ International Journal on Information and Communication Technology (IJoICT) Vol. 8 No. 1 (2022): June 2022 ]]>
Publisher : <![CDATA[School of Computing, Telkom University ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.21108/ijoict.v8i1.622
<![CDATA[Delivery of justice with the help of artificial intelligence is a current research interest. Machine learning with natural language processing (NLP) can classify the types of sexual harassment experiences into quid pro quo (QPQ) and hostile work environments (HWE). However, imbalanced data are often present in classes of sexual harassment classification on specific datasets. Data imbalance can cause a decrease in the classifier's performance because it usually tends to choose the majority class. This study proposes the implementation and performance evaluation of the synthetic minority over-sampling technique (SMOTE) to improve the QPQ and HWE harassment classifications in the sexual harassment experience dataset. The term frequency-inverse document frequency (TF-IDF) method applies document weighting in the classification process. Then, we compare naïve Bayes with K-Nearest Neighbor (KNN) in classifying sexual harassment experiences. The comparison shows that the performance of the naïve Bayes classifier is superior to the KNN classifier in classifying QPQ and HWE, with AUC values of 0.95 versus 0.92, respectively. The evaluation results show that by applying the SMOTE method to the naïve Bayes classifier, the precision of the minority class can increase from 74% to 90%.]]>
<![CDATA[Stock Portfolio Optimization on JII Index using Multi-Objective Mean-Absolute Deviation-Entropy ]]>
<![CDATA[Deni Saepudin]]>;
<![CDATA[Dimas Rizqi Guintana]]>
<![CDATA[ International Journal on Information and Communication Technology (IJoICT) Vol. 8 No. 1 (2022): June 2022 ]]>
Publisher : <![CDATA[School of Computing, Telkom University ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.21108/ijoict.v8i1.623
<![CDATA[Stock portfolio optimization is allocating stock assets from investors to manage return and risk. Investors need a high return portfolio with a given level of risk, and portfolio optimization can help to find the feasible one. The data used for this problem are stocks listed on the Jakarta Islamic Index (JII). The portfolio optimization methods are applied Mean-Absolute Deviation (MAD) and Entropy. MAD is used because it can solve the portfolio optimization problem for the nonnormal distribution of data. Meanwhile, entropy is used because it can better diversify the weight of stocks in the MAD portfolio. Experiment results in this study show that MAD-Entropy and Equal Weight portfolio outperform the MAD portfolio in Sharpe Ratio and Performance Ratio. MAD only excels in one period, influenced by a stock that has a fantastic return in a certain period.]]>
<![CDATA[General Depression Detection Analysis Using IndoBERT Method ]]>
<![CDATA[Ilham Rizki Hidayat]]>;
<![CDATA[Warih Maharani]]>
<![CDATA[ International Journal on Information and Communication Technology (IJoICT) Vol. 8 No. 1 (2022): June 2022 ]]>
Publisher : <![CDATA[School of Computing, Telkom University ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.21108/ijoict.v8i1.634
<![CDATA[Many of the tweets we discover on Twitter are concerning feelings of depression which will be caused by varied things. The amount of tweets additionally continues to increase. To be able to decide however depressed a user is, analysing tweets from users can facilitate with that. The method of analysing the detection of depression can help to supply applicable treatment for users who are detected to own depression. During this paper, the users to be analysed are users who have more than 1000 tweets and are Indonesian tweets. Then, crawling / retrieval of user tweet data is carried out. After that, data pre-processing is done. Once that done, using the IndoBERT method to classify the data obtained. In the end, this paper provides the accuracy value of this detection analysis using the IndoBERT method with an accuracy value of 51% and F1-Score of 31%.]]>
<![CDATA[Classification Analysis using CNN and LSTM on Wheezing Sounds ]]>
<![CDATA[Gustav Bagus Samanta]]>
<![CDATA[ International Journal on Information and Communication Technology (IJoICT) Vol. 8 No. 1 (2022): June 2022 ]]>
Publisher : <![CDATA[School of Computing, Telkom University ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.21108/ijoict.v8i1.621
<![CDATA[Asthma is a public health problem in almost all countries in the world. One of the symptoms that exist in asthmatics is wheezing. In several studies, wheezing has been classified using classification algorithm. However, the implemented classification algorithm still has a low level of accuracy. This study aims to determine the accuracy of the results from wheezing classification of respiratory sounds by comparing the algorithm.]]>
<![CDATA[Electronic Money Transactions Forecasting with Support Vector Regression (SVR) and Vector Autoregressive Moving Average (VARMA) ]]>
<![CDATA[I Nengah Dharma Pradnyandita]]>
<![CDATA[ International Journal on Information and Communication Technology (IJoICT) Vol. 8 No. 1 (2022): June 2022 ]]>
Publisher : <![CDATA[School of Computing, Telkom University ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.21108/ijoict.v8i1.632
<![CDATA[In today's digital era, the trend of payments with electronic money is rising. Some people have switched to do their way to the modern method such as electronic money. This is to improve the efficiency of the financial system. However, with the convenience and speed provided, if the use of electronic money is not being controlled properly, this can cause an unmanageable price of goods. In the context of controlling the risk of the use of electronic money, it is required to predict the use of electronic money in Indonesia. This paper, by using multivariate data analysis with the variable of electronic money transaction and Money supply (M1) as supporting variables in order to predict the nominal of electronic money transactions. The methods used are Vector Autoregressive Moving Average (VARMA) and Support Vector Regression (SVR). The results of the forecasting model will be compared using Mean Absolute Percentage Error (MAPE). According to the research that had been done, the SVR model had a better result compared to VARMA with a MAPE value of 3.577 %. This shows that the prediction data of the SVR model is close to actual data]]>
<![CDATA[Error Correction Codes Performance using Binary Phase Shift Keying over Fading Channel ]]>
<![CDATA[Hilal Hudan Nuha]]>;
<![CDATA[Abdi T. Abdalla]]>
<![CDATA[ International Journal on Information and Communication Technology (IJoICT) Vol. 8 No. 1 (2022): June 2022 ]]>
Publisher : <![CDATA[School of Computing, Telkom University ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.21108/ijoict.v8i1.646
<![CDATA[In a communication system, two main resources are used: transmission power and channel bandwidth. Transmission power is the average power of the transmitted signal. Channel bandwidth is defined as the frequency band allocated for the transmission of the message signal. A goal of general system design is to use these two resources as efficiently as possible. This scientific paper presents the experimental results of the Binary Phase Shift Keying (BSK) communication system on the additive white gaussian noise (AWGN) channel and the Fading channel. To improve system performance, error correction code (ECC) is used for encoding. ECC used include convolutional code (ConvCode) and Hamming code. Experimental results show that for BER=10^(-4) the coding gain of the ConvCode over Hamming code under AWGN is G=0.475dB. Whereas the coding gain of the ConvCode over unencoded BPSK is G=19.6dB.]]>
<![CDATA[Hybrid Hybrid wavelet and entropy features to monitor happy hypoxia based on photoplethysmogram signals ]]>
<![CDATA[Ayub Ginting]]>
<![CDATA[ International Journal on Information and Communication Technology (IJoICT) Vol. 8 No. 2 (2022): December 2022 ]]>
Publisher : <![CDATA[School of Computing, Telkom University ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.21108/ijoict.v8i2.629
<![CDATA[Happy hypoxia is a condition where patients experience decreasing oxygen saturation in their brains. In worst cases, Happy hypoxia can reduce the patient's consciousness and even death. Covid-19 has increased cases of happy hypoxia. Several studies have been conducted to detect the happy hypoxia. Existing research projects generally use photo plethysmography signals. However, the results show that the accuracy of happy hypoxia detection is still low. This study provides a solution to the above problems, by proposing a happy hypoxia detection system based on entropy and Discrete Wavele Transform (DWT) features that are combined with a classifier based on K Nearest Neighbor (KNN). The method used in this research is as below Hybrid Wavelet and Entropy Features method.Experiments on the proposed system have been carried out using data on Covid-19 patients from Haji Adam Malik Hospital in Medan.The experimental results show that the system proposed has an accuracy of 87%, sensitivity of 90% and specificity of 85]]>
<![CDATA[Recommender System Based on Matrix Factorization on Twitter Using Random Forest (Case Study: Movies on Netflix) ]]>
<![CDATA[Bagas Teguh Imani]]>;
<![CDATA[Erwin Budi Setiawan]]>
<![CDATA[ International Journal on Information and Communication Technology (IJoICT) Vol. 8 No. 2 (2022): December 2022 ]]>
Publisher : <![CDATA[School of Computing, Telkom University ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.21108/ijoict.v8i2.655
<![CDATA[In this day and age, there is a lot of entertainment that can be done, one of which is watching movies using the Netflix platform. When you want to watch, sometimes users can be confused about which movies to watch according to their tastes and interests, which requires a solution, namely by using a recommendation system. The recommendation system is a system that emerged as a solution to provide information by learning data from users with previously stored data items. One of the recommendation system techniques is Collaborative Filtering. By using Collaborative Filtering, this study will focus on using Matrix Factorization-based because it is considered more efficient and allows the incorporation of additional information in the data. This study will use the Random Forest algorithm to improve the results of good predictions. In this study, a recommendation system based on Matrix Factorization on Twitter will be made using Random Forest in a case study of films on Netflix. The experimental results have shown that the use of the system gets a Mean Absolute Error (MAE) value of 0.7641 to 0.8496 and a Root mean squared error (RMSE) of 1.0359 to 1.1935.]]>
<![CDATA[The Analysis of Support Vector Machine (SVM) on Monthly Covid-19 Case Classification ]]>
<![CDATA[Rifaldo Sitepu]]>
<![CDATA[ International Journal on Information and Communication Technology (IJoICT) Vol. 8 No. 2 (2022): December 2022 ]]>
Publisher : <![CDATA[School of Computing, Telkom University ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.21108/ijoict.v8i2.671
<![CDATA[Covid-19 is disease caused by the new corona virus called Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2). The effect of this virus usually causes infection on respiratory system. Covid-19 was rapidly spread globally. Experts said that the factor that caused this to spread rapidly is human mobility. Therefore, several countries create new rules so that it can suppress the spreading of this disease, by prohibiting a large scale gathering, keeping away distance with each other, mandatory rule of using mask, and the prohibition for the entry of their country. This research proposes a performance analysis of Support Vector Machine (SVM) to classify the monthly data of covid-19. The data used in this research is a series of covid-19 data of towns in Bandung from November 2020 until December 2021. From conducting this research It is found that the best accuracy was found on December 2021 with the accuracy of 100%, followed by July and August with the accuracy of 97%, and October with the value of 90%. We can conclude that Support Vector Machine (SVM), is a good method on classifying the monthly covid-19 data.]]>
