Science in Information Technology Letters

Journal Information
EISSN : 2722-4139
Published by: ASCEE Publications (10.31763)
Total articles ≅ 10
Filter:

Latest articles in this journal

Aram Ter-Sarkisov
Science in Information Technology Letters, Volume 1, pp 92-99; https://doi.org/10.31763/sitech.v1i2.202

Abstract:
This paper compares the models for the detection and segmentation of Ground Glass Opacity and Consolidation in chest CT scans. These lesion areas are often associated both with common pneumonia and COVID-19. We train a Mask R-CNN model to segment these areas with high accuracy using three approaches: merging masks for these lesions into one, deleting the mask for Consolidation, and using both masks separately. The best model achieves the mean average precision of 44.68% using MS COCO criterion on the segmentation across all accuracy thresholds. The classification model, COVID-CT-Mask-Net, learns to predict the presence of COVID-19 vs. common pneumonia vs. control. The model achieves the 93.88% COVID-19 sensitivity, 95.64% overall accuracy, 95.06% common pneumonia sensitivity, and 96.91% true-negative rate on the COVIDx-CT test split (21192 CT scans) using a small fraction of the training data. We also analyze the effect of the Non-Maximum Suppression of overlapping object predictions, both on the segmentation and classification accuracy. The full source code, models, and pre-trained weights are available on https://github.com/AlexTS1980/COVID-CT-Mask-Net.
N.M Saravana Kumar, K Hariprasath, N Kaviyavarshini, A Kavinya
Science in Information Technology Letters, Volume 1, pp 52-59; https://doi.org/10.31763/sitech.v1i2.167

Abstract:
Data mining is an in-depth study of enormous amounts of data present in an organization or institution’s repository. Business experts mostly utilize data analytics approaches to confirm their opinion. It will rapidly boost the global interest of the organization. In this scenario, the information and conclusion are gathered from Data analysis by data analytics. The experts also use it to validate, diagnose, or authenticate speculate layouts suppositions and completion of the analysis. In this paper, the prediction is based on grocery data sets by inspecting and analyzing the big mart sales data set. Among several predictive algorithms, data mining algorithms are considered for prediction. It includes Decision Tree, Naïve Bayes, Adaboost with Particle Swarm optimization, and Random forest. The proposed method of this research is a novel Naïve Bayes with a PSO algorithm. This algorithm optimizes the model iteratively. Exploration of the data must be done before prediction. The root means squared error (RMSE) is used as evaluation metrics for comparing the data mining algorithms. The proposed algorithm performs well and gives a lower RMSE value. So, the proposed algorithm fits the best model when compared with the existing algorithms. This paper describes the prediction of high-quality data analysis data and determines the efficiency of data mining algorithms.
Maniru Malami Umar, Aminu Mohammed, Abdulhakeem AbdulAzeez, Solomon Ordeun Yese
Science in Information Technology Letters, Volume 1, pp 60-72; https://doi.org/10.31763/sitech.v1i2.182

Abstract:
Call admission control (CAC) is a process of accepting new calls or handoff calls in a network while regulating the Quality of Service (QoS) of existing or active calls without degrading any call drop. CAC is an RRM technique and directly impacts QoS for individual connection and the overall system efficiency. This paper presents a short review of some existing CAC schemes proposed for the 3GPP LTE network. The review aims to guide researchers to know what CAC is, how it operates, and its benefits to the overall system performance. The schemes reviewed can be classified based on the aim they were proposed, such as guaranteeing QoS of calls, increasing the utilization of available network resources, reducing call blocking, and call dropping probabilities. The paper further summarised all reviewed schemes by highlighting each scheme's operations, strengths, and weaknesses.
, Jamal Mahmood, Mahmood Alfathe, Mohammed A. M. Abdullah
Science in Information Technology Letters, Volume 1, pp 73-82; https://doi.org/10.31763/sitech.v1i2.168

Abstract:
Nowadays, operation management performed by electronic systems in different life sectors is distributed and expanded swiftly due to several advantages that can be achieved, such as time reduction process, simplicity, and high accuracy of running operations. In the educational sector, performing student feedback electrolytically (using PC, Tablet, or Mobile) rather than the ordinary paper-based methods saves staff and students' time and makes managing the data easier. This paper proposes a new approach for the Electronic Student Feedback Management System based on web programming, which consists of input, process, store, and retrieves information using a database. The system is based on the Accreditation Board for Engineering and Technology (ABET) questionnaire model, which is implemented using web development tools (HTML, CSS, Javascript, MySQL, and PHP with Apache web-server). As a result, the proposed system was implemented successfully. The Electronic Engineering College officially adopted it at Ninevah University to perform the student feedback for the academic year 2018-2019, in which 4,282 student feedback applications (SFA) have been recorded. After that, statistical operations were done for extracting useful information effortlessly and accurately. As a result, more than 800 students made 4282 records have participated in the proposed system. This information can be quickly recorded and utilized to identify the weaknesses to address them in the next academic year.
Meiga Isyatan Mardiyah, Tuti Purwaningsih
Science in Information Technology Letters, Volume 1, pp 83-91; https://doi.org/10.31763/sitech.v1i2.160

Abstract:
Indonesia is an agricultural country with a variety of natural resources such as agriculture and plantations. Agriculture and plantations in Indonesia are diverse, such as rice fields that can produce rice, soybeans, corn, tubers, and others. Meanwhile, plantations in Indonesia are like forests with timber products, bamboo, eucalyptus oil, rattan, and others. However, rice fields, which are examples of agriculture, and forests that are examples of plantations, have the same characteristics. It is not easy to distinguish when viewed using aerial photographs or photographs taken from a certain height. For recognizing with certainty the shape of rice fields and forests when viewed using aerial photographs, it is necessary to establish a model that can accurately recognize the shape of rice fields and forest forms. A model is to utilize computational science to take information from digital images to recognize objects automatically. One method of deep learning that is currently developing is a Convolutional Neural Network (CNN). The CNN method enters (input data) in the form of an image or image. This method has a particular layer called the convulsive layer wherein an input image layer (input image) will produce a pattern of several parts of the image, which will be easier to classify later. The convolution layer has the function of learning images to be more efficient to be implemented. Therefore, researchers want to utilize this CNN method to classify forests and rice fields to distinguish the characteristics of forests and rice fields. Based on the classification results obtained by testing the accuracy of 90%. It can be concluded that the CNN method can classify images of forests and rice fields correctly.
Arief Bramanto Wicaksono Putra, Supriadi Supriadi, Aji Prasetya Wibawa, , Achmad Fanany Onnilita Gaffar
Science in Information Technology Letters, Volume 1, pp 9-16; https://doi.org/10.31763/sitech.v1i1.17

Abstract:
The unique features of an image can be obtained by changing the gray level by modifying the dynamic range of the gray level. The gray-level dynamic range modification technique is one technique to minimize the selected features. Bit rate reduction uses coding information with fewer bits than the original image (image compression). This study using the dynamic level of the gray level of a modified image with the concept of binary bit representation or also called bit manipulation. Using some binary bit representation options used: 4, 5, 6, and 7 of bit can obtain the best compression performance. Measurement of compression ratio and decompression error ratio to a benchmark comparison called compression performance, which is the ultimate achievement of this study. The results of this study show the use of 6-bit binary representation has the best performance, and the resulting image compression does not resize the resolution of the original image only visually looks different.The unique features of an image can be obtained by changing the gray level by modifying the dynamic range of the gray level. The gray-level dynamic range modification technique is one technique to minimize the selected features. Bit rate reduction uses coding information with fewer bits than the original image (image compression). This study using the dynamic level of the gray level of a modified image with the concept of binary bit representation or also called bit manipulation. Using some binary bit representation options used: 4, 5, 6, and 7 of bit can obtain the best compression performance.Measurement of compression ratio and decompression error ratio to a benchmark comparison called compression performance, which is the ultimate achievement of this study. The results of this study show the use of 6-bit binary representation has the best performance, and the resulting image compression does not resize the resolution of the original image only visually looks different.
Abdulrazak Yahya Saleh, Ermawatih Liansitim
Science in Information Technology Letters, Volume 1, pp 1-8; https://doi.org/10.31763/sitech.v1i1.1

Abstract:
Deep Convolutional Neural Networks (CNNs) have been established as a dominant class of models for image classification problems. This study aims to apply and analyses the accuracy of deep learning for classifying ripes on palm oil fruit. The CNN used to classify 628 images into 2 different classes. Furthermore, the experiment of CNN with 5 epochs gives promising classification results with an accuracy of 98%, which is better than previous methods. To sum up, this study was successfully solving an image classification by detected and differentiated the ripeness of oil palm fruit.
Pragathi Penikalapati, A Nagaraja Rao
Science in Information Technology Letters, Volume 1, pp 24-39; https://doi.org/10.31763/sitech.v1i1.32

Abstract:
Precise prediction of chronic diseases is the very basis of all healthcare informatics. Early diagnosis of the disease is crucial in delivering any healthcare service. The modern times witness our general vulnerability to several health disorders due to a stressful lifestyle causing anxiety and depression, or susceptibility to hypertension and diabetics or major diseases such as cancer or cardiovascular ailments. Hence, we should undergo periodic screening and diagnostic tests for such possible disorders to lead healthy lives. In this context, Machine Learning technology can play a pivotal role in developing Electronic Health Records (EHR) for implementing quick and comprehensively automated procedures in disease detection among the at-risk individuals at an early stage, so that accelerated processes of referral, counseling, and treatment can be initiated. The scope of the current paper is to survey the utilization of feature selection and techniques of Machine Learning, such as Classification and Clustering in the specific context of disease diagnosis and early prediction. This paper purposes of identifying the best models of Machine Learning duly supported by their performance indices, utility aspects, constraints, and critical issues in the specific context of their effective application in healthcare analytics for the benefit of practitioners and researchers.
Hartono Hartono, Erianto Ongko
Science in Information Technology Letters, Volume 1, pp 40-51; https://doi.org/10.31763/sitech.v1i1.34

Abstract:
Problems of Class Imbalance in data classification have received attention from many researchers. It is because the imbalance class will affect the accuracy of the classification results. The problem of the imbalance class itself will ignore the minority class, which is a class with a smaller number of instances even though the minority class is an exciting class to observe. In overcoming the imbalanced class problem, it is necessary to pay attention to diversity data, the number of classifiers, and also classification performance. Several methods have been proposed to overcome the imbalanced class problem, one of which is the Hybrid Approach Redefinition Method. This method is a good hybrid ensemble method in dealing with imbalance class problems, which can provide useful diversity data and also a smaller number of classifiers. This research will combine the Hybrid Approach Redefinition by replacing the use of SMOTE Boost by using Progressive Boosting to get better data diversity, a small number of classifiers, and better performance. This study will conduct testing in handling imbalance class problems using datasets sourced from the KEEL-Dataset Repository. The results of this study indicate that the Hybrid Approach Redefinition with Progressive Boosting will provide better results in the number of classifiers, data diversity, and classification performance.
Abdullah Ahmad Zarir, Saad Bashar, Amelia Ritahani Ismail
Science in Information Technology Letters, Volume 1, pp 17-23; https://doi.org/10.31763/sitech.v1i1.31

Abstract:
Generating natural language descriptions of the content of an image automatically is a complex task. Though it comes naturally to humans, it is not the same when making a machine do the same. But undoubtedly, achieving this feature would remarkably change how machines interact with us. Recent advancement in object recognition from images has led to the model of captioning images based on the relation between the objects in it. In this research project, we are demonstrating the latest technology and algorithms for automated caption generation of images using deep neural networks. This model of generating a caption follows an encoder-decoder strategy inspired by the language-translation model based on Recurrent Neural Networks (RNN). The language translation model uses RNN for both encoding and decoding, whereas this model uses a Convolutional Neural Networks (CNN) for encoding and an RNN for decoding. This combination of neural networks is more suited for generating a caption from an image. The model takes in an image as input and produces an ordered sequence of words, which is the caption.
Back to Top Top