http://repository.must.ac.ke/handle/123456789/1446 2026-04-13T14:56:40Z 2026-04-13T14:56:40Z Micheni, Fridah Kainyu http://repository.must.ac.ke/handle/123456789/1475 2025-04-15T08:16:40Z 2024-01-01T00:00:00Z

A Hybrid recommender model for career pathway selection in Competency-Based Education Micheni, Fridah Kainyu This study addresses the problem of inadequate guidance for learners in Competency-Based Education (CBE) when selecting career pathways, as current models often overlook important factors such as academic performance, personal interests, extracurricular activities, and career goals. In CBE, learners follow personalized, flexible learning paths based on their prior knowledge and skills, but career pathway decisions are frequently influenced by parents, teachers, and career counselors, missing the critical elements that help learners make informed choices. While recommender models are widely used in education for course selection and career advising, they have typically failed to integrate these diverse factors comprehensively. To address this gap, the study developed a hybrid recommender model designed to enhance career pathway selection in CBE. Using a mixed-method research design, data was collected through an online survey of 1,487 teachers from junior secondary schools in Meru County, focusing on factors influencing career pathway decisions. Analysis done in SPSS revealed that academic performance, personal interests, extracurricular activities, career goals, and job market trends are crucial to these decisions. Based on these insights, a CBE senior school dataset was created, and a hybrid recommender model was developed using hybrid filtering, deep neural networks, and random forest algorithms, combined through a stacking ensemble method. The model was validated using k-fold cross-validation and achieved an accuracy of 90.06% when applied to STEM career pathway tracks. These findings suggest that the hybrid model is effective in guiding learners toward appropriate STEM career pathway tracks in CBE. Future work could explore more advanced algorithms and expand the model to include additional career pathways.

2024-01-01T00:00:00Z Mutwiri, Lisper Kendi http://repository.must.ac.ke/handle/123456789/1463 2025-04-11T10:50:03Z 2023-01-01T00:00:00Z

Internet of Things Based Model for Hydropower Monitoring Mutwiri, Lisper Kendi In the energy sector, hydropower energy is very significant as it contributes to being a major source of renewable energy. Therefore, knowledge about hydropower energy and its existing challenges has led to an emerging need to obtain real-time data and a consistent monitoring model of the applications. This is meant to improve the performance and accuracy of the real- time data monitored by the model. In Kenya, a lack of hydrological datasets has been documented as a challenge in the Energy Act of 2018. This is primarily caused by an unprecedented reduction in the water levels in the hydropower plants, which leads to downtime and blackouts caused by little or no hydropower production. This study, therefore, sought to design, develop, and implement an Internet of Things-based model for hydropower monitoring. To achieve this objective, the study identified specific hydropower plants that are currently in operation, where data would be collected for validation, and the hardware and software to be used. In addition, the study also sought to identify an appropriate cloud storage service for storing the data set. The developed model was tested and validated with a total of 120 readings collected from the database. The selected site for data collection was Wanjii hydropower station, based in Murang'a County. The study used latency, throughput, consistency, and accuracy as metrics to evaluate the performance of the model. The T test was used to determine the significance of performance metrics. The study found that the monitoring model using LORA (long range) was feasible and practical during the testing and performed as expected during its validation. Based on the findings, the study recommends that the approach be scaled up and adopted for the entire hydropower system, including the mechanical valves. This would be more effective as its low-power and cheaper to embrace.

2023-01-01T00:00:00Z Muthee, George Mutwiri http://repository.must.ac.ke/handle/123456789/1462 2025-04-11T10:49:59Z 2023-01-01T00:00:00Z

A finite state transducer for morphological segmentation of Swahili verbs Muthee, George Mutwiri Morphological segmentation is a subtask of natural language processing (NLP) that specializes in identifying the constituent morphemes of words in a language. As a subtask of morphological analysis, morphological segmentation is a crucial preprocessing step that improves the overall output of the NLP system. Low resource languages have recently received attention in research, with researchers aiming to improve NLP in these languages. However, the finer details within languages are often overlooked, which has led to low quality results among a variety of studies. Since morphological analysis is an important task in NLP, extracting the morphological syntax of verbs thus remains crucial. In this work, the researcher demonstrates a finite state transducer for the morphological segmentation of the Swahili verb. This research work set out to achieve four objectives, namely to analyze key parameters for the morphological segmentation techniques of Swahili verbs, to implement a web scraper to populate a dataset of Swahili verbs, to integrate morphological segmentation parameters into a finite state transducer for Swahili verb segmentation, and to validate the finite state transducer on the Swahili verb. The model performs morphological analysis of the Swahili verb by identifying morphological slots such as the subject, object, derivational suffixes, and any grammatical errors within the verb. It was implemented as a finite state network built out of regular expressions in an object-oriented programming (OOP) language. The same finite state transducer was also implemented in the Xerox Finite State Tools (XFST). Input verbs were extracted from an online dictionary using a web scrapper and separated into two datasets. Dataset A comprised 163 simple Swahili verbs while dataset B comprised 715 non-Arabic verbs. The OOP model outperformed its XFST counterpart, achieving a 98.77% accuracy on dataset A and 68.67% accuracy on dataset B. The results from the experiments prove that OOP rule-based techniques perform better than their XFST-based counterparts. The research work was quantitative, with the accuracy of the models evaluated using experiments. This work is beneficial in optimizing search engines that use Swahili, where verbal keywords need to be segmented to obtain their root. This work is also pivotal in assisting learners new to Swahili in understanding the structure of the verb and enabling them to explore possible combinations of morphemes that make up a correctly formed verb. Further, the work significantly contributes towards the development of a spell checker, a corpus and a syntax analyzer for Swahili verbs.

2023-01-01T00:00:00Z