SU+ Digital Repository

SU+ is an online repository for the preservation and promotion of assorted digital content at Strathmore University

Off-Campus Access to restriced resources (including the ExamsBank) now requires registration using an @strathmore.edu email address

Authentication is NOT required for On-Campus Access to content

Photo by @inspiredimages

Communities in DSpace

Select a community to browse its collections.

Now showing 1 - 5 of 7

Conferences / Workshops / Seminars + 474
Documents and Proceedings of Conferences, Seminars, Workshops (and more) held at Strathmore University
Digital Archives 178
Assorted collections of resources covering various subject themes contributed by Faculty and Library Staff
Reports / Policies + 8
Public reports and policy documents
Research / Researchers / Publications 4332
Researcher Profiles / Conference presentations / Published research articles / Faculty and Corporate research outputs
Strathmore Heritage Collection 41
A digital chronicle of the History of the University presented through a mix of pictures, videos and digitized publications

Recent Submissions

Classification of anaemia types using supervised machine learning techniques
(Strathmore University, 2025) Onwong'a, C. K.
Anaemia reduction is one of the World Health assembly goals for 2025. Given the complex aetiology of anaemia, classification of nutritional anaemia using traditional methods has limitations and drawbacks. Traditional methods of classification rely heavily on analysis of complete blood count tests which need specialists and trained personnel, and present potential for errors in analysis. These traditional methods are also expensive and time consuming given the wait time between testing and getting the results. Machine learning based algorithms offer more accuracy and efficiency in the classification of anaemia given their ability to learn data and identify patterns. This study aimed at building a classification model for classifying nutritional types of anaemia using supervised machine learning techniques. The dataset that was utilized in this study was retrieved from Kaggle, an opensource dataset repository and used in accordance with the Open Database license. The dataset contained complete blood count test results for patients with proven cases of nutritional anaemia. The data was pre-processed and explored in preparation for model building. The features were all used in the model development because all the variables are different, and they contribute to the classification of anaemia. The models that were built are Naïve Bayes, random forest, XG Boost, decision trees, and multilayer perceptron. These models were tested using the testing set and their performances compared to find the better performing one. Hyperparameter tuning was done on some of the poorer performing models to try and improve their performance. The best performing model was the XG Boost classifier which achieved an accuracy of 98.85%. The poorest performing model was the Gaussian Naive Bayes model with an accuracy score of 0.7872. The SVM model was very computationally heavy and could not build. For deeper analysis of the model, metrics like recall, precision and F1 scores were measured. The XG boost model was then loaded to an interface for functionality testing. The tool was able to classify nutritional classes of anaemia based on complete blood count data entered by a user. This tool could potentially be plugged into hospitals and clinics to aid in the early detection, diagnosis and treatment by reducing the wait time between getting tested and getting results. This can be considered one of many steps towards anaemia reduction. Keywords: Anaemia Classification, Nutritional Anaemia classification, Supervised learning, SMOTE, XG Boost, Random Forest, Naive Bayes, Decision Trees, machine learning, multilayer perceptron.
PesaLink based mobile payment application for person to business payments
(Strathmore University, 2025) Mwalagho, M. W.
High transaction costs have continued to be of great concern for mobile payment users. The current setup of mobile payment is the provision of a wallet that can be topped up by either exchanging cash for electronic equivalent through an agent or by moving funds from a bank account into the wallet via a mobile banking application or internet banking. There is a cost incurred when moving money to a mobile wallet from a bank account to facilitate paying for utilities, goods and services. There is also a transaction fee charged by the Mobile Money Operator, this along with the bank transfer fees and taxation increases the cost of Mobile money payment transactions. The objective of this study was to research the current payment systems that are being used to make payments for goods, services and utilities along with their deficiencies, and propose an alternative solution using real-time bank-to-bank transfers through PesaLink. The focus was on enabling low-cost person-to-business payments without relying on mobile wallets, though the approach can be extended to person-to-person transactions as well. A prototype mobile payment application, developed using Agile methodology, demonstrated the feasibility of using PesaLink to eliminate the need for wallet-based transfers. Testing with sandbox bank accounts confirmed successful real-time fund transfers, reduced transaction steps, and cost-saving potential. The iterative Agile process allowed flexibility in development, accommodating changes in requirements during the project lifecycle. Keywords: Mobile payments, real-time payments, person-to-business payments, low-cost, no wallet, Agile methodology
A Mobile application based system for tomato pest and disease detection
(Strathmore University, 2025) Kamau, R. N.
Precision farming is an approach to farming that uses technology and data to reduce farming costs, improve crop yields, and optimize resource utilization. It relies on a range of tools and technologies to collect, process, and analyse data to derive actionable information that is used to make informed decisions about planting, disease and pest control, and harvesting. Tools used in data collection include Internet of Things (IoT) devices, Unmanned Aerial Vehicles (UAVs), Geographic Information Systems (GIS), and weather stations. Data analytics, GIS mapping and machine learning are used to process, analyse, and make sense of the data. Several factors limit smallholder farmers from adopting precision farming solutions. These include high cost of implementation, limited digital infrastructure such as the Internet, lack of awareness and limited knowledge in utilising these solutions. This research aimed to bridge this gap by developing a system that uses machine learning to detect tomato pests and diseases and suggest treatment and prevention measures through the mobile phone. A review of the factors influencing agricultural productivity among smallholder farmers, and existing solutions that attempt to mitigate these factors was carried out. A requirement analysis was carried out to establish the viability of developing a mobile application-based system for tomato disease detection. Subsequently, this dissertation developed a deep learning Convolutional Neural Network (CNN) model trained using 16,880 images of tomato diseases. The trained model achieved a validation accuracy of 98.22%. A mobile application was developed, and the machine learning model embedded into it. Prototyping software development methodology was adopted to develop the system. The system was tested at several stages, to ensure that it met the set requirements for performance and functional requirements. The system was able to detect tomato diseases and provide the user with disease treatment and prevention recommendations. Keywords: precision farming, machine learning, convolutional neural network, smallholder farming.
A Predictive model for blood demand and supply: a case of obstetric emergencies at Kenyatta National Hospital
(Strathmore University, 2025) Mutai, K. V.
Maternal health is still a very critical concern in Kenya, with obstetric emergencies posing considerable challenges to maternal health and mothers’ well-being. One key factor that contributes to maternal mortality is the unavailability of blood during obstetric complications, which is essential for saving lives during emergencies related to childbirth. Unfortunately, many lives have been lost under the circumstances due to a lack of timely information regarding blood requests and donations. These delays in finding blood to manage such situations more effectively cause loss of life. The study aimed to develop a predictive web-based model to monitor blood demand and supply and ensure blood is available to save the lives of mothers and newborn babies during obstetric emergencies using Kenyatta National Hospital (KNH) as a case study. This will be achieved through improved blood bank management, monitoring blood availability and providing timely information to blood stakeholders, minimizing delays in blood donation and supply requests during obstetric emergencies. This will also help reduce maternal and infant mortality cases, thus improving maternal health outcomes. If successful, the proposed solution will go a long way in improving blood donation practices in the healthcare system. The model leveraged modern application development technologies and tools, including Predictive Analytics, Artificial Intelligence, and Machine Learning libraries, to forecast blood demand and supply based on historical data. The research adopts an experimental design to gain in-depth insights into the current practices, challenges, and patterns in blood management. This approach facilitated the identification of critical variables influencing blood demand and supply, leveraging both qualitative and quantitative data to make correct blood demand versus supply predictions. By exploring existing systems, healthcare workflows, and historical data, the study will provide a foundation for constructing an effective predictive model tailored to the unique requirements of obstetric care. Keywords: Obstetric Emergency, Maternal Health, Blood Donation, Predictive Analytics.
Automated event attendance recording tool using facial recognition
(Strathmore University, 2025) Makhatsa, B. S.
Monitoring and measuring participant attendance at events is crucial for several reasons. In addition to offering organizers and sponsors insights into an event's popularity, it is essential for assessing the economic, social, and environmental impacts of the event. Various methods for recording attendance are currently in use, including manual methods such as signing a register which are often inefficient and prone to errors. While electronic attendance systems, such as NFC cards and fingerprint readers, exist, NFC cards are vulnerable to impersonation, and fingerprint readers may raise health concerns due to contact with shared devices. In contrast, facial recognition technology offers a promising solution for improving the accuracy of attendance records at events while mitigating hygiene risks associated with fingerprint readers. However, facial recognition systems tend to be computationally intensive. This research developed an automated event attendance recording tool that integrates Hash-Based Indexing with facial recognition algorithms to enhance the accuracy of attendance records while reducing computational load through record indexing. The study employed agile methodology for developing the automated attendance recording tool. Testing and evaluation utilized publicly available image databases to train the machine learning image recognition model and assess its performance. The primary evaluation metrics included the accuracy of image identification and the duration of transaction processing. The model achieved 95% accuracy in face recognition, with its performance further analysed using the confusion matrix and classification report. The developed tool provided an interface for event organizers to create events and record attendance offering utilizing the full capabilities of CNN for image recognition. The developed tool offered an interface that allowed event organizers to create events and record attendance, fully utilizing the capabilities of CNN for image recognition and hash-based indexing for faster retrieval of records. Keywords: Attendance, Agile methodology, Electronic attendance systems, Facial recognition, Fingerprint readers, Machine learning, Hash-Based Indexing, Record indexing