Print ISSN: 1812-125X

Online ISSN: 2664-2530

Main Subjects : Artificial Intelligent


Medical Images Classification Using Artificial Intelligence

Tasneem Mustafa; Jamal Salahaldeen Alneamy

Journal of Education and Science, In Press
DOI: 10.33899/edusj.2022.133358.1224

ABSTRACT:
In recent years, the use of computing has increased along with medical skills, and this had impressive results in terms of classification and treatment, in addition to facilitating the matter of medical personnel. This was evident during the Corona pandemic, which infected millions around the world, which. There is an urgent need for software tools to help classify the disease without the need to resort to doctors. The matter is not limited to the classification of corona disease, but it also extends to the expansion of the discovery of other diseases such as malaria, skin cancer and other diseases that afflict large numbers of people. Malaria is an infectious disease caused by the Plasmodium parasite, and according to some statistics, the total number of infections in 2019 reached about 228 million cases around the world. As for skin cancer, it is considered one of the serious diseases that affect humans because the skin plays a key role in protecting muscles and bones, and therefore cancer will affect all body functions.
CNN have made great strides in many intractable problems in image processing and classification, but their performance depends on their hyperparameters which is a tedious task if done manually. Therefore, experts in the field of deep learning aspire to improve its performance sometimes by integrating it with other algorithms such as Particle Swarm Optimization, Gray Wolf Optimization, Genetic Algorithm or firefly. All of these algorithms gave different results than the others, that is, they gave different levels of performance.

Classification of Software Systems attributes based on quality factors using linguistic knowledge and machine learning: A review.

Abdulrhman Ali; Nada Nimat Saleem

Journal of Education and Science, In Press
DOI: 10.33899/edusj.2022.134024.1245

Both The functionality and the non-functionality for what the system does as well as doesn't of software systems requirements are documented in a Software Requirements Specification (SRS).
Moreover in requirements engineering, system requirements classify into several categories such as functional, quality and constraint classes.
Therefore, we evaluate several machine learning approaches as well as methodologies that mentioned in previous literature in term of automatic requirements extraction, then the classification based on methodically reviewing for many previous works on software requirements classification to assist software engineers in selecting the best requirement classification technique. Therefore, the study aim to get answer for several questions that related to: What machine learning algorithms were used for the classification process of the requirements, How these algorithms work and how they're evaluated, What methods were used for extracting features from a text, What evaluation criteria were used in comparing results and What machine learning techniques and methods were provided the highest accuracy.

Implementation of OCR using Convolutional Neural Network (CNN): A Survey

Ahmed Abdulrahman Alkaddo; Dujan Albaqal

Journal of Education and Science, In Press
DOI: 10.33899/edusj.2022.133711.1236

Recently, character recognition and deep learning have caught the attention of many researchers. Optical Character Recognition (OCR) usually takes an image of the character as input and generates the identical character as output. The important role that OCR does is to transform printed materials into digital text files. Convolutional Neural Network (CNN) is an influential model that is generous with bright results in optical character recognition (OCR). The state-of-the-art performance which exists in deep neural networks is usually used to handle frequently recognition and classification problems. Many applications are using it, for instance, robotics, traffic monitoring, articles digitization, etc. CNN is designed to adaptively and automatically learn features by using many kinds of layers (convolution layers, pooling layers, and fully connected layers). In this paper we will go through the advantages and recent usage of CNN in OCR and why it’s important to use it in handwritten and printed text recognition and what subjects we can use this technique for. Researchers are progressively using CNN for the machine-printed characters and recognition of handwritten, that is because CNN architectures are suitable for recognition tasks by inputting some images

A Review Of Clustering Methods Based on Artificial Intelligent Techniques

Baydaa ibraheem Khaleel

Journal of Education and Science, 2022, Volume 31, Issue 2, Pages 69-82
DOI: 10.33899/edusj.2022.133092.1218

Due to the development in various areas of life, the development of the Internet, and the presence of many dataset, and in order to obtain useful information from the rapidly increasing volumes of digital data, there must be theories and computational tools to help humans extract the useful information they need from this data. Large data is collected from many different services and resources. Clustering is one of the most basic and well-known methods of data mining and extraction and obtaining useful information. The technique of recognizing natural groups or clusters within several datasets based on some measure of similarity is known as data clustering. Many researchers have introduced and developed many clustering algorithms based on the different methods of artificial intelligence techniques. Finding the right algorithms greatly helps in organizing information and extracting the correct answer from different database queries. This paper provides an overview of the different clustering methods using artificial intelligence and finding the appropriate clustering algorithm to process different data sets. We highlight the best-performing clustering algorithm that gives effective and correct clustering for each data set.

Software Development Effort Estimation Techniques: A Survey

farah basil alhamdany; laheeb Mohammad Ibrahim

Journal of Education and Science, 2022, Volume 31, Issue 1, Pages 80-92
DOI: 10.33899/edusj.2022.132274.1201

Software Effort Estimation (SEE) is used in accurately predicting the effort in terms of (person–hours or person–months). Although there are many models, Software Effort Estimation (SEE) is one of the most difficult tasks for successful software development. Several SEE models have been proposed. However, software effort overestimation or underestimation can lead to failure or cancellation of a project.
Hence, the main target of this research is to find a performance model for estimating the software effort through conduction empirical comparisons using various Machine Learning (ML) algorithms. Various ML techniques have been used with seven datasets used for Effort Estimation. These datasets are China, Albrecht, Maxwell, Desharnais, Kemerer, Cocomo81, Kitchenham, to determine the best performance for Software Development Effort Estimation. Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and R-Squared were the evaluation metrics considered. Results and experiments with various ML algorithms for software effort estimation have shown that the LASSO algorithm with China dataset produced the best performance compared to the other algorithms.

Intelligence System for Multi-Language Recognition

Fawziya Ramo; Mohammed Naif Kannah

Journal of Education and Science, 2022, Volume 31, Issue 1, Pages 93-110
DOI: 10.33899/edusj.2022.132223.1200

Language classification systems are used to classify spoken language from a particular phoneme sample and are usually the first step of many spoken language processing tasks, such as automatic speech recognition (ASR) systems Without automatic language detection, spoken speech cannot be properly analyzed and grammar rules cannot be applied, causing failures Subsequent speech recognition steps. We propose a language classification system that solves the problem in the image field, rather than the sound field. This research identified and implemented several low-level features using Mel Frequency Cepstral Coefficients, which extract traits from speech files of four languages (Arabic, English, French, Kurdish) from the database (M2L_Dataset) as the data source used in this research.
A Convolutional Neuron Network is used to operate on spectrogram images of the available audio snippets. In extensive experiments, we showed that our model is applicable to a range of noisy scenarios and can easily be extended to previously unknown languages, while maintaining classification accuracy. We released our own code and extensive training package for language classification systems for the community.
CNN algorithm was applied in this research to classify and the result was perfect, as the classification accuracy reached 97% between two languages if the sample length was only one second, but if the sample length was two seconds, the classification accuracy reached 98%. While the classification among three languages, the classification accuracy reached 95% if the sample length was only one second, but if the sample length was two seconds, the classification accuracy reached 96%.

Detection of citrus diseases using a fuzzy neural network

Huda Saad Taher; Baydaa I. Khaleel

Journal of Education and Science, 2021, Volume 30, Issue 5, Pages 125-135
DOI: 10.33899/edusj.2021.130928.1179

The objective is to use AI techniques to build a citrus image recognition system and to produce an integrated program that will assist plant protection professionals in determining whether the disease is infected and early detection for the purpose of taking the necessary preventive measures and reducing its spread to other plants. In this research, the RBF and FRBF networks were used and applied to 830 images, to detect whether citrus fruits were healthy or ill. At first, the preprocessing of these images was done, and they were reduced to 250 x 250 pixels, and the features were extracted from them using the co-occurrence matrix method (GLCM) after setting the gray level at 8 gradients and 1 pixel distance, 21 statistical features were derived, and then these features were introduced to RBF after determine the number of input layer nodes by 21 , 20 for the hidden layer and 1 node for output layer, the centers were randomly selected from the training data and the weights were also randomly selected and trained using the Pseudo Inverse method. The RBF network was hybridized with the fuzzy logic using the FCM method, the fuzziness parameter = 2.3 was selected, and a new network called FRBF was acquired. These networks were trained and tested in training data (660 images) and testing (170 images) for citrus fruits. The detection rate was then calculated, and the results showed that the (FRBF) had a higher accuracy of 98.24% compared to RBF of 94.71%.

Ransomware Detection System Based on Machine Learning

Omar Shamil Ahmed; Omar Abdulmunem Ibrahim Al-Dabbagh

Journal of Education and Science, 2021, Volume 30, Issue 5, Pages 86-102
DOI: 10.33899/edusj.2021.130760.1173

In every day, there is a great growth of the Internet and smart devices connected to the network. On the other hand, there is an increasing in number of malwares that attacks networks, devices, systems and apps. One of the biggest threats and newest attacks in cybersecurity is Ransom Software (Ransomware). Although there is a lot of research on detecting malware using machine learning (ML), only a few focuses on ML-based ransomware detection. Especially attacks targeting smartphone operating systems (e.g., Android) and applications. In this research, a new system was proposed to protect smartphones from malicious apps through monitoring network traffic. Six ML methods (Random Forest (RF), k-Nearest Neighbors (k-NN), Multi-Layer Perceptron (MLP), Decision tree (DT), Logistic Regression (LR), and eXtreme Gradient Boosting (XGB)) are applied on CICAndMal2017 dataset which consists of benign and various kinds of android malware samples. A 603288 benign and ransomware samples were extracted from this collection. Ransomware samples are collected from 10 different families. Several types of feature selection techniques have been used on the dataset. Finally, seven performance metrics were used to determine the best one of feature selection and ML classifiers for ransomware detection. The experiments results imply that DT and XGB outperforms other classifiers with best detection accuracy are more than (99.30%) and (99.20%) for (DT) and (XGB) respectively.

Human Activity Recognition: literature Review

mais irreem atheed; Dena Rafaa Ahmed; Rashad Adhed Kamal

Journal of Education and Science, 2021, Volume 30, Issue 5, Pages 12-29
DOI: 10.33899/edusj.2021.130293.1162

Abstract:
Human activity recognition has an important role in the interaction between human and human relationships because it provides information about a person's identity, personality, activities, psychological state, and health, all this information is difficult to extract due to the difficulty of a person's ability to identify the activities of another person and is considered one of the basic research topics in the scientific fields in the field of computer vision and machine learning. the purpose of human activity recognition (HAR) is to identify the different human activities throw monitoring and register the human activates and the various surrounded environment, by using computers, the human activity recognition researches which depending on visions is the basics of lots of applications even video monitoring or health care and security monitoring and the interaction between the human and the computers.
In this research, a review of the newest development in the human activity recognition branch have been studied, and the different ways to recognize the human actions, an important detail have been shown to preview the HAR researches and the methodologies used to represent the human activates and its classifications, to provide an overview of the HAR methods and comparing them

Text dependent speaker identification system based on deep learning

Qasim Sadiq Mahmood; Yusra Faisal Al-Irahyim

Journal of Education and Science, 2021, Volume 30, Issue 4, Pages 141-160
DOI: 10.33899/edusj.2021.130144.1161

Speaker identification techniques are one of those most advanced modern technologies and there are many different systems had been developed, from methods that used to extract characteristics and classification. The applications of Speech identification are quite difficult and requires modern technologies with a large number of audio samples and resources.
In this research, the system of speaker identification had been designed based on a text (the word or sentences are pre-defined) which give the system the capability to identify the speaker in the least time, number of training samples and resources. The system consists four main parts, the first one is to create audio databases. In the study, two audio databases were relied upon, the first being a database (QS- Dataset) and the second database (audioMNIST_meta). The databases were processed and configured in a way that was explained in the body of the research later. The second part of the research is to extract the characteristics through the pitch coefficients algorithm, while the third part is the use of the neural network as a classifier. And the last part of the research is to verify the work and results of the system.
The test results showed the ability of the MNN network to deal with the smallest number of data, as it achieved a percentage of 100%. As for large data, it ranged from 80% to 81%. Unlike CNN network, the results were not good for the few data, from 60% to 76%, and with large data it was The results are excellent, from 91% to 96%.

Speaker Recognition: Progression and challenges

Yusra Faisal Al-Irahyim; Qasim Sadiq Mahmood

Journal of Education and Science, 2021, Volume 30, Issue 4, Pages 59-68
DOI: 10.33899/edusj.2021.129802.1150

Speaker recognition is one of the field topics widely used in the field of speech technology, many research works has been conducted and little progress has been made in the past five to six years, and due to the advancement of deep learning techniques in most areas of machine learning, it has been replaced previous research methods in speaking recognition and verification. The topic of deep learning is now the most advanced solution to verifying and identifying a speaker's identity. The algorithms used are (x-vectors) and (i-vectors) which are considered the baseline in modern work. The aim of this study is to review deep learning methods applied in identifying speakers and tasks for validating older solutions (Gaussian mixture model, Gaussian mixture super vector model and i-vector model) to new solutions using deep neural networks (deep belief network, deep corrective learning network). ) As well as the types of metrics to verify the speaker (cosine distance, probabilistic linear discrimination analysis) as well as the databases used for neural network training (TIMIT, VCTK, VoxCeleb2, LibriSpeech).