Abstract: In this work, a methodology is presented for identifying the optimal digitization parameters for the analog signal of ultrasonic sensors. These digitization parameters are the resolution of the analog to digital conversion and the sampling rate. This is accomplished though the derivation of characteristic curves based on Fano inequality and the calculation of the mutual information content over a given dataset. The mutual information is calculated between the examples in the dataset and the corresponding variation in the feature that needs to be estimated. The optimal parameters are identified in a manner that ensures optimal estimation performance while preventing inefficiency in using unnecessarily powerful analog to digital converters.
Abstract: This paper aims to provide an interpretation of artificial neural networks (ANNs) and explore some of its implications. The interpretation views ANNs as a memory which encodes instances of experience. An experiment explores the behavior of encoding and retrieval of instances from memory. A localised representation ANN is created that allows control over encoding and retrieved memory sample size and is experimented with using the MNIST digits dataset. The relationship between input familiarity, conflict within retrieved samples, and error rates is described and demonstrated to be an effective driver for memory encoding. Results indicate that selective encoding and retrieval samples that allow detection of memory conflicts produce optimal performance, and that error rates are normally distributed with input familiarity and conflict. By using input familiarity and sample consistency to guide memory encoding, the number of encoding trials on the dataset were reduced to 18.33% of the training data while maintaining good recognition performance on the test data.
Abstract: Twitter is one of the most popular social media platforms where users share their opinions on different subjects. Twitter can be considered a great source for mining text due to the high volumes of data generated through the platform daily. Many industries such as telecommunication companies can leverage the availability of Twitter data to better understand their markets and make an appropriate business decision. This study performs topic modeling on Twitter data using Latent Dirichlet Allocation (LDA). The obtained results are benchmarked with another topic modeling technique, Latent Semantic Indexing (LSI). The study aims to retrieve topics on a Twitter dataset containing user tweets on South African Telcos. Results from this study show that LSI is much faster than LDA. However, LDA yields better results with higher topic coherence by 8% for the best-performing model in this experiment. A higher topic coherence score indicates better performance of the model.
Abstract: The use of technology has benefited society in more ways than one ever thought possible. Unfortunately, as society’s knowledge of technology has advanced, so has its knowledge of ways to use technology to manipulate others. This has led to a simultaneous advancement in the world of fraud. Machine learning techniques can offer a possible solution to help decrease these advancements. This research explores how the use of various machine learning techniques can aid in detecting fraudulent activity across two different types of fraudulent datasets, and the accuracy, precision, recall, and F1 were recorded for each method. Each machine learning model was also tested across five different training and testing splits in order to discover which split and technique would lead to the most optimal results.
Abstract: Digital investigators often have a hard time spotting evidence in digital information. It has become hard to determine which source of proof relates to a specific investigation. A growing concern is that the various processes, technology, and specific procedures used in the digital investigation are not keeping up with criminal developments. Therefore, criminals are taking advantage of these weaknesses to commit further crimes. In digital forensics investigations, artificial intelligence (AI) is invaluable in identifying crime. Providing objective data and conducting an assessment is the goal of digital forensics and digital investigation, which will assist in developing a plausible theory that can be presented as evidence in court. This research paper aims at developing a multiagent framework for digital investigations using specific intelligent software agents (ISAs). The agents communicate to address particular tasks jointly and keep the same objectives in mind during each task. The rules and knowledge contained within each agent are dependent on the investigation type. A criminal investigation is classified quickly and efficiently using the case-based reasoning (CBR) technique. The proposed framework development is implemented using the Java Agent Development Framework, Eclipse, Postgres repository, and a rule engine for agent reasoning. The proposed framework was tested using the Lone Wolf image files and datasets. Experiments were conducted using various sets of ISAs and VMs. There was a significant reduction in the time taken for the Hash Set Agent to execute. As a result of loading the agents, 5% of the time was lost, as the File Path Agent prescribed deleting 1,510, while the Timeline Agent found multiple executable files. In comparison, the integrity check carried out on the Lone Wolf image file using a digital forensic tool kit took approximately 48 minutes (2,880 ms), whereas the MADIK framework accomplished this in 16 minutes (960 ms). The framework is integrated with Python, allowing for further integration of other digital forensic tools, such as AccessData Forensic Toolkit (FTK), Wireshark, Volatility, and Scapy.
Abstract: In developing countries, most roads in rural areas are dirt road. They require frequent maintenance since they are affected by erosive events, such as rain or wind, and the transit of heavy-weight trucks and machinery. Early detection of damages on the road condition is a key aspect, since it allows to reduce the maintenance time and cost, and also the limitations for other vehicles to travel through. Most proposals that help address this problem require the explicit participation of drivers, a permanent internet connection, or important instrumentation in vehicles or roads. These constraints limit the suitability of these proposals when applied into developing regions, like Latin America. This paper proposes an alternative method, based on unattended crowdsensing, to determine the quality of dirt roads in rural areas. This method involves the use of a mobile application that complements the road condition surveys carried out by organizations in charge of the road network maintenance, giving them early warnings about road areas that could be requiring maintenance. Drivers can also take advantage of the early warnings while they move through these roads. The method was evaluated using information from a public dataset. Although they are preliminary, the results indicate the proposal is potentially suitable to provide awareness about dirt roads condition to drivers, transportation authority and road maintenance companies.
Abstract: Face recognition is the problem of identifying or recognizing individuals in an image. This paper investigates a possible method to bring a solution to this problem. The method proposes an amalgamation of Principal Component Analysis (PCA), K-Means clustering, and Convolutional Neural Network (CNN) for a face recognition system. It is trained and evaluated using the ORL dataset. This dataset consists of 400 different faces with 40 classes of 10 face images per class. Firstly, PCA enabled the usage of a smaller network. This reduces the training time of the CNN. Thus, we get rid of the redundancy and preserve the variance with a smaller number of coefficients. Secondly, the K-Means clustering model is trained using the compressed PCA obtained data which select the K-Means clustering centers with better characteristics. Lastly, the K-Means characteristics or features are an initial value of the CNN and act as input data. The accuracy and the performance of the proposed method were tested in comparison to other Face Recognition (FR) techniques namely PCA, Support Vector Machine (SVM), as well as K-Nearest Neighbour (kNN). During experimentation, the accuracy and the performance of our suggested method after 90 epochs achieved the highest performance: 99% accuracy F1-Score, 99% precision, and 99% recall in 463.934 seconds. It outperformed the PCA that obtained 97% and KNN with 84% during the conducted experiments. Therefore, this method proved to be efficient in identifying faces in the images.
Abstract: Time Series Forecasting (TSF) is used to predict the target variables at a future time point based on the learning from previous time points. To keep the problem tractable, learning methods use data from a fixed length window in the past as an explicit input. In this paper, we study how the performance of predictive models change as a function of different look-back window sizes and different amounts of time to predict into the future. We also consider the performance of the recent attention-based transformer models, which had good success in the image processing and natural language processing domains. In all, we compare four different deep learning methods (Recurrent Neural Network (RNN), Long Short-term Memory (LSTM), Gated Recurrent Units (GRU), and Transformer) along with a baseline method. The dataset (hourly) we used is the Beijing Air Quality Dataset from the website of University of California, Irvine (UCI), which includes a multivariate time series of many factors measured on an hourly basis for a period of 5 years (2010-14). For each model, we also report on the relationship between the performance and the look-back window sizes and the number of predicted time points into the future. Our experiments suggest that Transformer models have the best performance with the lowest Mean Absolute Errors (MAE = 14.599, 23.273) and Root Mean Square Errors (RSME = 23.573, 38.131) for most of our single-step and multi-steps predictions. The best size for the look-back window to predict 1 hour into the future appears to be one day, while 2 or 4 days perform the best to predict 3 hours into the future.
Abstract: Arabic Sentiment Analysis (SA) is one of the most common research fields with many open areas. This paper proposes different pre-processing steps and a modified methodology to improve the accuracy using normal Support Vector Machine (SVM) classification. The paper works on two datasets, Arabic Sentiment Tweets Dataset (ASTD) and Extended Arabic Tweets Sentiment Dataset (Extended-ATSD), which are publicly available for academic use. The results show that the classification accuracy approaches 86%.
Abstract: With the advent of three-dimension (3D) technology, there are lots of research in converting 2D images to 3D images. The main difference between 2D and 3D is the visual illusion of depth in 3D images. In the recent era, there are more depth estimation techniques. The objective of this paper is to convert 2D images to 3D images with less computation time. For this, the input image is divided into blocks from which the depth information is obtained. Having the depth information, a depth map is generated. Then the 3D image is warped using the original image and the depth map. The proposed method is tested on Make3D dataset and NYU-V2 dataset. The experimental results are compared with other recent methods. The proposed method proved to work with less computation time and good accuracy.
Abstract: Due to the quick increase in the volume of Arabic opinions posted on various social media, Arabic sentiment analysis has become one of the most important areas of research. Compared to English, there is very little works on Arabic sentiment analysis, in particular aspect-based sentiment analysis (ABSA). In ABSA, aspect extraction is the most important task. In this paper, we propose a semantic ABSA approach for standard Arabic reviews to extract explicit aspect terms and identify the polarity of the extracted aspects. The proposed approach was evaluated using HAAD datasets. Experiments showed that the proposed approach achieved a good level of performance compared with baseline results. The F-measure was improved by 19% for the aspect term extraction tasks and 55% aspect term polarity task.
Abstract: Multiple classifier systems combine several individual classifiers to deliver a final classification decision. However, an increasingly controversial question is whether such systems can outperform the single best classifier, and if so, what form of multiple classifiers system yields the most significant benefit. Also, multi-target tracking detection using multiple sensors is an important research field in mobile techniques and military applications. In this paper, several multiple classifiers systems are evaluated in terms of their ability to predict a system’s failure or success for multi-sensor target tracking tasks. The Bristol Eden project dataset is utilised for this task. Experimental and simulation results show that the human activity identification system can fulfil requirements of target tracking due to improved sensors classification performances with multiple classifier systems constructed using boosting achieving higher accuracy rates.
Abstract: This study demonstrates an alternative stochastic imputation approach for large datasets when preferred commercial packages struggle to iterate due to numerical problems. A large country conflict dataset motivates the search to impute missing values well over a common threshold of 20% missingness. The methodology capitalizes on correlation while using model residuals to provide the uncertainty in estimating unknown values. Examination of the methodology provides insight toward choosing linear or nonlinear modeling terms. Static tolerances common in most packages are replaced with tailorable tolerances that exploit residuals to fit each data element. The methodology evaluation includes observing computation time, model fit, and the comparison of known values to replaced values created through imputation. Overall, the country conflict dataset illustrates promise with modeling first-order interactions, while presenting a need for further refinement that mimics predictive mean matching.
Abstract: Electroencephalography (EEG) investigations of the brain computer interfaces are based on the electrical signals resulting from neural activities in the brain. In this paper, it is offered a method for classifying motor imagery EEG signals. The suggested method classifies EEG signals into two classes using the wave atom transform, and the transform coefficients are assessed, creating the feature set. Classification is done with SVM and k-NN algorithms with and without feature selection. For feature selection t-test approaches are utilized. A test of the approach is performed on the BCI competition III dataset IIIa.
Abstract: The goal of motor image brain computer interface research is to create a link between the central nervous system and a computer or device. The most important signal for brain-computer interface is the electroencephalogram. The aim of this research is to explore a set of effective features from EEG signals, separated into frequency bands, using wave atom sub-bands to discriminate right and left-hand motor imagery signals. Over the transform coefficients, feature vectors are constructed for each frequency range and each transform sub-band, and their classification performances are tested. The method is validated using EEG signals from the BCI competition III dataset IIIa and classifiers such as support vector machine and k-nearest neighbors.
Abstract: Aflatoxins are highly poisonous and carcinogenic compounds produced by species of the genus Aspergillus spp. that can infect a variety of agricultural foods, including dried figs. Biological and environmental factors, such as population, pathogenicity and aflatoxinogenic capacity of the strains, topography, soil and climate parameters of the fig orchards are believed to have a strong effect on aflatoxin levels. Existing methods for aflatoxin detection and measurement, such as high-performance liquid chromatography (HPLC), and enzyme-linked immunosorbent assay (ELISA), can provide accurate results, but the procedures are usually time-consuming, sample-destructive and expensive. Predicting aflatoxin levels prior to crop harvest is useful for minimizing the health and financial impact of a contaminated crop. Consequently, there is interest in developing a tool that predicts aflatoxin levels based on topography and soil analysis data of fig orchards. This paper describes the development of a risk assessment tool for the contamination of aflatoxin on dried figs, based on the location and altitude of the fig orchards, the population of the fungus Aspergillus spp. in the soil, and soil parameters such as pH, saturation percentage (SP), electrical conductivity (EC), organic matter, particle size analysis (sand, silt, clay), concentration of the exchangeable cations (Ca, Mg, K, Na), extractable P and trace of elements (B, Fe, Mn, Zn and Cu), by employing machine learning methods. In particular, our proposed method integrates three machine learning techniques i.e., dimensionality reduction on the original dataset (Principal Component Analysis), metric learning (Mahalanobis Metric for Clustering) and K-nearest Neighbors learning algorithm (KNN), into an enhanced model, with mean performance equal to 85% by terms of the Pearson Correlation Coefficient (PCC) between observed and predicted values.
Abstract: Biometric identification is to identify unique features in a person like fingerprints, iris, ear, and voice recognition that need the subject's permission and physical contact. Gait biometric is used to identify the unique gait of the person by extracting moving features. The main advantage of gait biometric to identify the gait of a person at a distance, without any physical contact. In this work, the gait biometric is used for person re-identification. The person walking naturally compared with the same person walking with bag, coat and case recorded using long wave infrared, short wave infrared, medium wave infrared and visible cameras. The videos are recorded in rural and in urban environments. The pre-processing technique includes human identified using You Only Look Once, background subtraction, silhouettes extraction and synthesis Gait Entropy Image by averaging the silhouettes. The moving features are extracted from the Gait Entropy Energy Image. The extracted features are dimensionality reduced by the Principal Component Analysis and recognized using different classifiers. The comparative results with the different classifier show that Linear Discriminant Analysis outperform other classifiers with 95.8% for visible in the rural dataset and 94.8% for longwave infrared in the urban dataset.
Abstract: The primary tool currently used to pre-process 10X chromium single-cell ATAC-seq data is Cell Ranger, which can take very long to run on standard datasets. To facilitate rapid pre-processing that enables reproducible workflows, we present a suite of tools called scATAK for pre-processing single-cell ATAC-seq data that is 15 to 18 times faster than Cell Ranger on mouse and human samples. Our tool can also calculate chromatin interaction potential matrices and generate open chromatin signal and interaction traces for cell groups. We use scATAK tool to explore the chromatin regulatory landscape of a healthy adult human brain and unveil cell-type specific features, and show that it provides a convenient and computational efficient approach for pre-processing single-cell ATAC-seq data.
Abstract: Identification of plant diseases has been performed using machine learning and deep learning models on the datasets containing images of healthy and diseased plant leaves. The current study carries out an evaluation of some of the deep learning models based on convolutional neural network architectures for identification of plant diseases. For this purpose, the publicly available New Plant Diseases Dataset, an augmented version of PlantVillage dataset, available on Kaggle platform, containing 87,900 images has been used. The dataset contained images of 26 diseases of 14 different plants and images of 12 healthy plants. The CNN models selected for the study presented in this paper are AlexNet, ZFNet, VGGNet (four models), GoogLeNet, and ResNet (three models). The selected models are trained using PyTorch, an open-source machine learning library, on Google Colaboratory. A comparative study has been carried out to analyze the high degree of accuracy achieved using these models. The highest test accuracy and F1-score of 99.59% and 0.996, respectively, were achieved by using GoogLeNet with Mini-batch momentum based gradient descent learning algorithm.
Abstract: This paper presents a software product quality evaluation model based on the ISO/IEC 25010 quality model. The evaluation characteristics and sub characteristics were identified from the ISO/IEC 25010 quality model. The multidimensional structure of the quality model is based on characteristics such as functional suitability, performance efficiency, compatibility, usability, reliability, security, maintainability, and portability, and associated sub characteristics. Random numbers are generated to establish the decision maker’s importance weights for each sub characteristics. Also, random numbers are generated to establish the decision matrix of the decision maker’s final scores for each software product against each sub characteristics. Thus, objective criteria importance weights and index scores for datasets were obtained from the random numbers. In the proposed model, five different software product quality evaluation datasets under three different weight vectors were applied to multiple criteria decision analysis method, preference analysis for reference ideal solution (PARIS) for comparison, and sensitivity analysis procedure. This study contributes to provide a better understanding of the application of MCDMA methods and ISO/IEC 25010 quality model guidelines in software product quality evaluation process.