Zambia ICT Journal

Utilizing Machine Learning for Accurate Property Valuation: A Regression Model Analysis

2025-01-02T15:13:48+00:00

This research investigated the application of machine learning (ML) regression models to improve property valuation accuracy, addressing limitations of traditional methods. The study applied Random Forest (RF) and Support Vector Regression (SVR) models to a dataset of 59,180 property records from the Luanshya Municipal Council. Key features such as LAND_VALUE, MARKET_VALUE, and IMPROVEMENT_VALUE were used as inputs. The models' performance was evaluated using the Data, Reasoning, and Usefulness (DRU) Evaluation Framework. Results showed that both RF and SVR outperformed traditional methods, with RF achieving an R² of 0.9995. Machine Learning models demonstrated potential for more accurate property valuations, enabling fairer tax assessments, reduced manual effort, and improved urban planning decisions. Future research should address data quality and model explainability challenges.

Integrating ICT Solutions for Sustainable Agriculture: Addressing Food Security in the Face of Climate Change

2025-01-02T14:49:46+00:00

Climate change poses an escalating threat to global food security, necessitating innovative approaches to agricultural practices. This paper explores how Information and Communication Technology (ICT) can revolutionize agriculture, emphasizing precision agriculture, mobile technologies, and data analytics. The study examines socio-economic and infrastructural barriers to ICT adoption, offering a detailed framework for scaling ICT solutions in rural and resource-constrained areas. By leveraging a comprehensive analysis of case studies and emerging technologies, the findings highlight the transformative potential of ICT in enhancing resource efficiency, increasing yields, and mitigating climate change impacts. This research contributes to bridging theory and practice, supporting sustainable agricultural development and global food security.

Artificial Intelligence-Driven Data Science for Enhancing TB Treatment Outcomes and Reducing Mortality Rates

2025-01-02T14:13:49+00:00

Tuberculosis (TB) remains a leading global health concern, ranking among the top causes of death worldwide. It surpasses HIV in mortality, with approximately 1.3 million deaths annually among HIV-negative individuals and 214,000 among those co-infected with HIV [1]. Despite progress in reducing TB mortality rates, as evidenced by Zambia’s decrease from 759 to 361 deaths per 100,000 population between 2000 and 2017, TB continues to be a significant health challenge in high-burden regions like southern Africa [1]. In Zambia's Itezhi-Tezhi district, factors contributing to TB mortality include advanced age, poor treatment adherence, extra-pulmonary TB, and complications related to co-infection with HIV. While systems such as SmartCare and YATHU DR TB have been developed to manage records for drug-susceptible and drug-resistant TB, there is a critical need for more advanced tools to further enhance treatment outcomes. This paper proposes an AI-driven automated system that utilizes data science techniques to improve TB treatment outcomes and reduce mortality rates. The system employs data mining to track and analyze comprehensive patient records throughout the treatment phases, including intensive and continuation phases. It gathers and evaluates data on patient demographics, drug adherence, treatment progress, and outcomes in comparison with similar cases that have achieved successful treatment. By leveraging AI algorithms to predict treatment outcomes based on historical data, healthcare providers can gain valuable insights into patient progress, enabling more timely and effective interventions. This proactive approach aims to address challenges in TB management, enhance patient monitoring, and ultimately improve overall treatment efficacy. The integration of AI and data science into TB care represents promising advancement in combating one of the world’s most persistent infectious diseases.

Investigation of the suitability of existing Maize Plant Leaf Disease detection and classification approaches: Challenges and Open Issues

2025-01-02T15:43:05+00:00

Maize, a crucial staple crop in Zambia and many other regions, is seriously threatened by several leaf diseases, such as Gray Leaf Spot, Maize Streak Virus, and Northern Corn Leaf Blight. Early detection and accurate classification of these diseases are challenging due to the time-consuming and error prone nature of traditional detection techniques, such as visual inspections by farmers or experts. In recent years, deep learning has shown promise as an automated method for identifying and categorizing plant diseases. This paper describes a deep learning-based framework for identifying leaf diseases in maize plants and identifies the main obstacles and unresolved problems in the field. The lack of large enough and diverse datasets is one of the main obstacles to using deep learning for maize disease diagnosis, particularly in certain regions such as Zambia. The usefulness of existing models is limited in real-world scenarios because they frequently fail to generalize across various environmental circumstances, such as variances in climate, illumination, and soil type. In addition, the class imbalance creates a big gap in datasets with overrepresentations of specific diseases, which skews model predictions. The lack of lightweight, deployable models appropriate for low-resource settings, including rural farms with limited access to high-end computing equipment, is another significant gap. Furthermore, deep learning models are frequently perceived as "black boxes," and because farmers and other agricultural specialists need explicable insights into disease forecasts, they are less likely to be adopted due to the lack of interpretability of the models. This paper addresses the requirement for ongoing model updates to deal with changing disease patterns and investigates the possibilities of domain adaptation and transfer learning approaches in enhancing model generalization across conditions and locations. The report concludes by urging a concentrated effort to incorporate regional farmers and agricultural stakeholders in the development process to guarantee that the solutions are workable, approachable, and contextually appropriate. Even though deep learning has a lot of promise to improve the detection of maize leaf disease, there are still several issues that need to be resolved to produce more scalable and successful solutions. This study identifies these gaps and makes recommendations for how to close them in the future to support food security and sustainable agricultural development in areas like Zambia.

Investigating the Factors Influencing Students’ Adoption of Generative AIs in Universities: A Case of the Copperbelt University

2025-01-02T15:19:21+00:00

Generative Artificial Intelligence technologies (GenAIs) have increasingly become essential components of students' learning practices in Universities, requiring an examination of the levels of acceptance and factors responsible for acceptance and usage. The study therefore investigated the levels of awareness and adoption and the factors influencing the adoption of Generative-AIs amongst University students, employing the Technology Acceptance Model (TAM) as a theoretical framework. Data from 285 students within the Copperbelt University at different levels of study was analyzed using SPSS. The instrument used in collecting the data was an online questionnaire. Results indicated there are high levels of awareness (88%) and adoption (82%) of GenAIs in learning by students in Universities and there was a relatively high usage frequency (51%). The research also revealed that Expected Benefits, Perceived Usefulness, Attitude Toward Technology, and Behavioral Intention all significantly impacted students’ adoption of Generative AI. This study underscores the need to promote a culture of adopting and integrating new promising innovations such ads GenAIs in Universities, and at the same time establish ethical guidelines to promote responsible GenAIs use within education. This research also contributes to the understanding of factors responsible for GenAI adoption in higher educational settings and helps inform strategies for equitable access and responsible innovation.

Integrating Chicken Fecal Image Analysis with Machine Learning for Early Detection of Poultry Diseases in Developing Countries

2025-01-02T14:59:53+00:00

The health and productivity of poultry farms are significantly impacted by the timely detection of diseases within chicken houses. Manual disease monitoring in poultry is laborious and prone to errors, underscoring the need for sustainable, efficient, reliable, and cost-effective farming practices. The adoption of advanced technologies, such as artificial intelligence (AI), is essential to address this need. Smart farming solutions, particularly machine learning, have proven to be effective predictive analytical tools for large volumes of data, finding applications in various domains including medicine, finance, and sports, and now increasingly in agriculture. Poultry diseases like Coccidiosis can lower chicken productivity if they are not detected early on. Machine learning, Deep learning algorithms can assist with the early identification of diseases. In this study, a Convolutional Neural Network based framework has been proposed to classify poultry diseases by distinguishing healthy and unhealthy fecal images. Unhealthy images can be a sign of poultry diseases. The Image Classification dataset was used to train a model, and it was discovered that it performed with an accuracy of 84.99%, 90.05% on the training set, testing set respectively. When the proposed network's performance was evaluated, it was discovered that the proposed model was unquestionably the best one for classifying chicken disease. This study explores the benefits of automated chicken disease detection as a function of smart farming in Zambia.

Open-Source Automatic Circuit Recloser with Remote Control and Monitoring

2025-01-02T14:27:43+00:00

According to statistics, a large amount of faults in transmission and distribution networks are temporary faults that disappear a certain time after de-energization of the faulted sections of the network and then power is restored manually after the fault has cleared. Most recently, technology advancement has made available Automatic Circuit Recloser (ACR). This device makes it possible to recover the original status of the network without any human interaction. Unfortunately, for Africa and Zambia in particular the ACR system is proprietary, expensive and need “experts” to install, maintain and repair it. This study uses an open source embedded prototype single-line-to-ground (SLG) system that uses fixed dead time to discriminate transient from permanent faults. Simultaneously, an SMS is sent to the technician or the control center. A program generated dead time it will attempt to switch ON for a number of given times beyond which the system is locked out until reset. The prototype was successfully simulated and implemented.

Exploring Cancer Risks Amidst Climate-Induced Drought and Using Artificial Intelligence to Enhance Cancer Risk Prevention Strategies

2025-01-02T15:29:31+00:00

The rising number of droughts due to climate change raises serious concerns for public health. This issue may lead to an increase in cancer cases. The study closely examines the relationship between prolonged droughts & cancer rates. It specifically takes into account poor water quality, elevated exposure to cancer-causing agents, along with challenges in agricultural systems. To clarify the connection between climate change-induced droughts and elevated cancer risk, artificial intelligence (AI) will be employed in this study. AI tools & models will assist in predicting the cancer risks associated with these droughts. Additionally, this research aims to enhance early detection methods & develop targeted prevention strategies. By integrating insights from climate science, public health, and Artificial Intelligence, the goal is to provide practical recommendations for mitigating and preventing cancer risks in regions severely impacted by drought. The primary objective here is to equip policymakers, healthcare professionals, & researchers with reliable strategies. Such strategies will bolster public health as we confront the realities of a shifting climate. This paper adopts an exploratory research design to delve into the risks of cancer linked to climate-induced droughts.

Leveraging Artificial Intelligence - Driven Automata for Improved Dropout Prediction in Schools

2025-01-02T15:06:39+00:00

This work investigates the use of Artificial Intelligence - driven automata to predict student dropout risks, emphasizing how artificial intelligence can enhance educational sustainability. By integrating automata theory with cutting-edge machine learning techniques, the study develops a predictive framework to identify students at high risk of dropping out. The research includes a thorough review of automata theory, Artificial Intelligence applications in education, and current dropout prevention strategies. Using automata-based algorithms, we analyze educational datasets to model student behavior and risk factors. Preliminary findings indicate that these models can accurately forecast dropout probabilities, facilitating timely interventions to boost student retention and promote a more equitable educational system. Additionally, this work aligns with broader sustainability goals by supporting improved educational outcomes, which are crucial for advancing green economies amidst climate change. The results highlight the potential of automata-based AI in dropout prevention and offer insights for future research on integrating AI solutions in educational settings to address climate-related challenges.

An Integrated NLP and Machine Learning Model for Detecting Smishing Attacks on Mobile Money Platforms

2025-01-02T14:38:59+00:00

The Southern African Development Community (SADC), notably Zambia, has experienced a rise in mobile financial services, which has increased vulnerability to SMS phishing attacks leading to financial losses which has had negative socio-economic effects. This paper presents the cybersecurity risks associated with SMS-phishing on mobile money platforms and proposes a detection model using machine learning (ML) and natural language processing (NLP). The model employs Random Forest and Naïve Bayes algorithms for classification, utilizing NLP techniques such as Named Entity Recognition and part-of-speech tagging to extract relevant text features from SMS messages. The model is trained on both real-world and synthetic SMS datasets consisting of Bemba and English, with performance evaluated using precision, recall, F1 score, and ROC curves. Initial results demonstrate high accuracy and effective detection capabilities. The paper also stresses the need for user education to complement the technological solution in enhancing mobile financial security.

Machine Translation for Improved Access to Healthcare Information in Remote Zambian Communities

2025-01-02T15:36:54+00:00

Language barriers pose a significant obstacle to achieving national health goals in Zambia. Patients struggle to understand critical medical information that in the process hinder diagnosis, treatment and preventive care. This research proposes a groundbreaking approach to bridge this gap by developing a comprehensive parallel medical linguistic corpora resource and a specialized Machine Translation (MT) system tailored to Zambia’s diverse languages. Zambia’s Eighth National Development Plan prioritizes improved healthcare outcomes. However, widespread language barriers between patients, healthcare providers and public health initiatives create communication gaps. Costly interpreter services are of ten unavailable, leaving vulnerable populations without access to crucial medical information. This is particularly concerning for mothers seeking maternal care, communities battling disease outbreaks and researchers struggling to include diverse populations in medical studies. This study will address these challenges by developing a next generation MT system and a parallel medical corpus specifically designed for Zambian Low Resourced Languages(LRL). Additionally, the study will pioneer new evaluation metrics tailored to the specific needs of medical translations. The MT system, accessible through a user friendly mobile application, will empower Zambians to access vital health information in their native languages. This will improve communication between patients and healthcare providers that will lead to better diagnoses, treatment plans and overall health outcomes. Furthermore, the research will contribute valuable insights to the broader field of MT, advancing the technology for low-resource languages and specialized domains worldwide.