Journal of Computing Theories and Applications

An Intelligent Telediagnosis of Acute Lymphoblastic Leukemia using Histopathological Deep Learning

2024-06-01T08:58:52+00:00

Leukemia, a global health challenge characterized by malignant blood cell proliferation, demands innovative diagnostic techniques due to its increasing incidence. Among leukemia types, Acute Lymphoblastic Leukemia (ALL) emerges as a particularly aggressive form affecting diverse age groups. This study proposes an advanced mechanized system utilizing Deep Neural Networks for detecting ALL blast cells in microscopic blood smear images. Achieving a remarkable accuracy of 97% using MobileNetV2, our system demonstrates high sensitivity and specificity in identifying multiple ALL sub-types. Furthermore, we introduce cutting-edge telediagnosis software facilitating real-time support for clinicians in promptly and accurately diagnosing various ALL subtypes from microscopic blood smear images. This research aims to enhance leukemia diagnosis efficiency, which is crucial for the timely intervention and managing this life-threatening condition.

Integrating Convolutional Neural Network and Weighted Moving Average for Enhanced Human Fall Detection Performance

2024-06-01T08:58:52+00:00

This study proposes an approach for human fall classification utilizing a combination of Weighted Moving Average (WMA) and Convolutional Neural Networks (CNN) on the SisFall dataset. Falls among elderly individuals pose a significant public health concern, necessitating effective automated detection systems for timely intervention and assistance. The SisFall dataset, comprising accelerometer data collected during simulated falls and activities of daily living, serves as the basis for training and evaluating the proposed classification system. The proposed method begins by preprocessing accelerometer data using a WMA technique to enhance signal quality and reduce noise. Subsequently, the preprocessed data are fed into a CNN architecture optimized for feature extraction and fall classification. The CNN leverages its ability to automatically learn discriminative features from raw sensor data, enabling robust and accurate classification of fall and non-fall events. Experimental results demonstrate the efficacy of the proposed approach in accurately distinguishing between fall and non-fall activities, achieving high classification performance metrics such as accuracy, precision, recall, and F1-score. Comparative analysis with existing methods showcases the WMA-CNN hybrid approach's superiority in classification accuracy and robustness. Overall, the proposed methodology presents a promising framework for real-time human fall classification using sensor data, offering potential applications in wearable devices, ambient assisted living systems, and healthcare monitoring technologies to enhance safety and well-being among elderly individuals.

Development of a Model to Classify Skin Diseases using Stacking Ensemble Machine Learning Techniques

2024-06-01T08:58:52+00:00

Skin diseases are highly prevalent and transmissible. It has been one of the major health problems that most people face. The diseases are dangerous to the skin and tend to spread over time. A patient can be cured of these skin diseases if they are detected on time and treated early. However, it is difficult to identify these diseases and provide the right medications. This study's research objectives involve developing an ensemble machine learning based model for classifying Erythemato-Squamous Diseases (ESD). The ensemble techniques combine five different classifiers, Naïve Bayes, Support Vector Classifier, Decision Tree, Random Forest, and Gradient Boosting, by merging their predictions and utilizing them as input features for a meta-classifier during training. We tested and validated the ensemble model using the dataset from the University of California, Irvine (UCI) repository to assess its effectiveness. The Individual classifiers achieved different accuracies: Naïve Bayes (85.41%), Support Vector Machine (98.61%), Random Forest (97.91%), Decision Tree (95.13%), Gradient Boosting (95.83%). The stacking method yielded a higher accuracy of 99.30% and a precision of 1.00, recall of 0.96, F1 score of 0.97, and specificity of 1.00 compared to the base models. The study confirms the effectiveness of ensemble learning techniques in classifying ESD.

Top-Heavy CapsNets Based on Spatiotemporal Non-Local for Action Recognition

2024-06-01T08:58:52+00:00

To effectively comprehend human actions, we have developed a Deep Neural Network (DNN) that utilizes inner spatiotemporal non-locality to capture meaningful semantic context for efficient action identification. This work introduces the Top-Heavy CapsNet as a novel approach for video analysis, incorporating a 3D Convolutional Neural Network (3DCNN) to apply the thematic actions of local classifiers for effective classification based on motion from the spatiotemporal context in videos. This DNN comprises multiple layers, including 3D Convolutional Neural Network (3DCNN), Spatial Depth-Based Non-Local (SBN) layer, and Deep Capsule (DCapsNet). Firstly, the 3DCNN extracts structured and semantic information from RGB and optical flow streams. Secondly, the SBN layer processes feature blocks with spatial depth to emphasize visually advantageous cues, potentially aiding in action differentiation. Finally, DCapsNet is more effective in exploiting vectorized prominent features to represent objects and various action features for the ultimate label determination. Experimental results demonstrate that the proposed DNN achieves an average accuracy of 97.6%, surpassing conventional DNNs on the traffic police dataset. Furthermore, the proposed DNN attains average accuracies of 98.3% and 80.7% on the UCF101 and HMDB51 datasets, respectively. This underscores the applicability of the proposed DNN for effectively recognizing diverse actions performed by subjects in videos.

Adversarial Convolutional Neural Network for Predicting Blood Clot Ischemic Stroke

2024-06-01T08:58:52+00:00

Digital Pathology Image Analysis (DPIA) is one of the areas where deep learning (DL) techniques offer modern, cutting-edge functionality. Convolutional Neural Network (CNN) technology outperforms the competition in classification, segmentation, and detection tasks while being just one of numerous DL techniques. Classification, segmentation, and detection methods can often be used to address DPIA concerns. Some difficulties can also be resolved using pre- and post-processing techniques. However, other CNN models have been investigated for use in addressing DPIA-related issues. Furthermore, the research seeks to explore how susceptible the model is to adversarial attacks and suggest strategies to counteract them. To predict ischemic strokes caused by blood clots, the authors of this study developed CNN with a pixel brightness transformation (PBT) technique for image enhancement and developed several approaches of image augmentation techniques to increase and provide the learning model with more diverse features. Also, adversarial training was integrated into CNN models to train the model with perturbed data in order to assess the impact of adversarial noise at different stages of training. Several metrics, including precision, F1-score, accuracy, and recall, are utilized to assess the experiments' effectiveness. The research findings indicate that employing transfer learning with a deep learning model achieved an accuracy of up to 97% using the ReLU activation function. Also, data augmentation helps improve the accuracy of the model.

Integrating Structural Causal Model Ontologies with LIME for Fair Machine Learning Explanations in Educational Admissions

2024-05-27T13:24:07+00:00

This study employed knowledge discovery in databases (KDD) to extract and discover knowledge from the Benue State Polytechnic (Benpoly) admission database and used a structural causal model (SCM) ontological framework to represent the admission process in the Nigerian polytechnic education system. The SCM ontology identified important causal relations in features needed to model the admission process and was validated using the conditional independence test (CIT) criteria. The SCM ontology was further employed to identify and constrain input features causing bias in the local interpretable model-agnostic explanations (LIME) framework applied to machine learning (ML) black-box predictions. The ablation process produced more stable LIME explanations devoid of fairness bias compared to LIME without ablation, with higher prediction accuracy (91% vs. 89%) and F1 scores (95% vs. 94%). The study also compared the performance of different ML models, including Gaussian Naïve Bayes, Decision Trees, and Logistic Regression, before and after ablation. The limitation is that the SCM ontology is qualitative and context-specific, so the fair-LIME framework can only be extrapolated to similar contexts. Future work could compare other explanation frameworks like Shapley on the same dataset. Overall, this study demonstrates a novel approach to enforcing fairness in ML explanations by integrating qualitative SCM ontologies with quantitative ML/LIME methods.

Effects of Data Resampling on Predicting Customer Churn via a Comparative Tree-based Random Forest and XGBoost

2024-06-02T20:49:42+00:00

Customer attrition has become the focus of many businesses today – since the online market space has continued to proffer customers, various choices and alternatives to goods, services, and products for their monies. Businesses must seek to improve value, meet customers' teething demands/needs, enhance their strategies toward customer retention, and better monetize. The study compares the effects of data resampling schemes on predicting customer churn for both Random Forest (RF) and XGBoost ensembles. Data resampling schemes used include: (a) default mode, (b) random-under-sampling RUS, (c) synthetic minority oversampling technique (SMOTE), and (d) SMOTE-edited nearest neighbor (SMOTEEN). Both tree-based ensembles were constructed and trained to assess how well they performed with the chi-square feature selection mode. The result shows that RF achieved F1 0.9898, Accuracy 0.9973, Precision 0.9457, and Recall 0.9698 for the default, RUS, SMOTE, and SMOTEEN resampling, respectively. Xgboost outperformed Random Forest with F1 0.9945, Accuracy 0.9984, Precision 0.9616, and Recall 0.9890 for the default, RUS, SMOTE, and SMOTEEN, respectively. Studies support that the use of SMOTEEN resampling outperforms other schemes; while, it attributed XGBoost enhanced performance to hyper-parameter tuning of its decision trees. Retention strategies of recency-frequency-monetization were used and have been found to curb churn and improve monetization policies that will place business managers ahead of the curve of churning by customers.