Browsing by Author "Nyambo, Devotha"

Now showing 1 - 20 of 33

Application of Multiple Unsupervised Models to Validate Clusters Robustness in Characterizing Smallholder Dairy Farmers
(Hindawi The Scientific World Journal, 2019-01-02) Nyambo, Devotha; Luhanga, Edith; Yonah, Zaipuna; Mujibi, Fidalis
The heterogeneity of smallholder dairy production systems complicates service provision, information sharing, and dissemination of new technologies, especially those needed to maximize productivity and profitability. In order to obtain homogenous groups within which interventions can bemade, it is necessary to define clusters of farmers who undertake similar management activities. This paper explores robustness of production cluster definition using various unsupervised learning algorithms to assess the best approach to define clusters. Data were collected from 8179 smallholder dairy farms in Ethiopia and Tanzania. From a total of 500 variables, selection of the 35 variables used in defining production clusters and household membership to these clusters was determined by Principal Component Analysis and domain expert knowledge. Three clustering algorithms, K-means, fuzzy, and Self-Organizing Maps (SOM), were compared in terms of their grouping consistency and prediction accuracy. The model with the least household reallocation between clusters for training and testing data was deemed the most robust. Prediction accuracy was obtained by fitting a model with fixed effects model including production clusters on milk yield, sales, and choice of breeding method. Results indicated that, for the Ethiopian dataset, clusters derived fromthe fuzzy algorithm had the highest predictive power (77% for milk yield and 48% for milk sales), while for the Tanzania data, clusters derived from Self-Organizing Maps were the best performing.The average cluster membership reallocation was 15%, 12%, and 34% for K-means, SOM, and fuzzy, respectively, for households in Ethiopia. Based on the divergent performance of the various algorithms evaluated, it is evident that, despite similar information being available for the study populations, the uniqueness of the data fromeach country provided an over-riding influence on cluster robustness and prediction accuracy.The results obtained in this study demonstrate the difficulty of generalizing model application and use across countries and production systems, despite seemingly similar information being collected.
An Approach for Systematically Analyzing and Specifying Security Requirements for the Converged Web-Mobile Applications
(Scientific Publishing Center, 2014-09-01) Nyambo, Devotha; Yonah, Zaipuna; Tarimo, Charles
As the use of web and mobile applications is becoming pervasive for service delivery and user mobility support, enterprises are now increasingly fighting against a huge number of emerging security threats which interfere with the process of service delivery. As an attempt to help the enterprises in dealing with the emerging security threats in the converged service delivery architecture, this paper presents a methodology for security threat analysis and security requirements specification in web/mobile applications development. The presented methodology is based on a case study Livestock Data Center (LDC) system, which is being developed and it allows both web and mobile interfaces as service delivery channels. Hence the system serves as a representative of other similar setups of service delivery. In addition to the processes of analysis and security specification, the methodology involves threat modeling as well. There are several threat models in the literature. The STRIDE threats model is one among the existing threats models that is used to identify security threats that needs to be addressed in systems such as the LDC system. The STRIDE threats model has been used to identify the likely security threats to our case study. On applying the STRIDE threats model the following threats were identified as prominent: sensitive data exposure, weak server side controls, client side injection, and weak authentication and authorization. The identified security threats were compared to existing threats in traditional web and mobile applications separately in order to figure out the changes when the two computing platforms come together. The findings from our case study have shown that the proposed methodology for security threat analysis and security design can be useful in security requirements specifications in the converged web-mobile applications during development, and can be generally used to assist developers of other similar systems.
Automated Optimization-Based Deep Learning Models for Image Classification Tasks
(Computers, 2023-09-01) Migayo, Daudi; Kaijage, Shubi; Swetala, Stephen; Nyambo, Devotha
Applying deep learning models requires design and optimization when solving multi- faceted artificial intelligence tasks. Optimization relies on human expertise and is achieved only with great exertion. The current literature concentrates on automating design; optimization needs more attention. Similarly, most existing optimization libraries focus on other machine learning tasks rather than image classification. For this reason, an automated optimization scheme of deep learning models for image classification tasks is proposed in this paper. A sequential-model-based optimization algorithm was used to implement the proposed method. Four deep learning models, a transformer-based model, and standard datasets for image classification challenges were employed in the experiments. Through empirical evaluations, this paper demonstrates that the proposed scheme improves the performance of deep learning models. Specifically, for a Virtual Geometry Group (VGG-16), accuracy was heightened from 0.937 to 0.983, signifying a 73% relative error rate drop within an hour of automated optimization. Similarly, training-related parameter values are proposed to improve the performance of deep learning models. The scheme can be extended to automate the optimization of transformer-based models. The insights from this study may assist efforts to provide full access to the building and optimization of DL models, even for amateurs.
Blockchain-based Data Storage Security Architecture for e-Health Care Systems: A Case of Government of Tanzania Hospital Management Information System
(IJCSNS, 2022-03) Mnyawi, Richard; Kombe, Cleverence; Sam, Anael; Nyambo, Devotha
Health information systems (HIS) are facing security challenges on data privacy and confidentiality. These challenges are based on centralized system architecture creating a target for malicious attacks. Blockchain technology has emerged as a trending technology with the potential to improve data security. Despite the effectiveness of this technology, still HIS are suffering from a lack of data privacy and confidentiality. This paper presents a blockchain-based data storage security architecture integrated with an e-Health care system to improve its security. The study employed a qualitative research method where data were collected using interviews and document analysis. Execute-order-validate Fabric’s storage security architecture was implemented through private data collection, which is the combination of the actual private data stored in a private state, and a hash of that private data to guarantee data privacy. The key findings of this research show that data privacy and confidentiality are attained through a private data policy. Network peers are decentralized with blockchain only for hash storage to avoid storage challenges. Cost-effectiveness is achieved through data storage within a database of a Hyperledger Fabric. The overall performance of Fabric is higher than Ethereum. Ethereum’s low performance is due to its execute-validate architecture which has high computation power with transaction inconsistencies. E-Health care system administrators should be trained and engaged with blockchain architectural designs for health data storage security. Health policymakers should be aware of blockchain technology and make use of the findings. The scientific contribution of this study is based on; cost-effectiveness of secured data storage, the use of hashes of network data stored in each node, and low energy consumption of Fabric leading to high performance
Brix and Alcohol Content Monitoring using Wireless Sensor Network
(International Journal of Advances in Scientific Research and Engineering, 2021-07) Willa, Victor; Nyambo, Devotha; Sam, Anael
Fermentation process plays an important role in the production of wine and beer, as these ferments convert Brix (sugar) into alcohol. Thus fermentation plays a key role in beer production hence monitoring entire activities is essential to breweries. As such, this project aimed to monitor the by-products of fermentation processes which are Brix (Sugar concentration) and Alcohol in order to improve the Quality of alcohol. Different stages of fermentation produce different Brix and alcohol percentages by volume. A system using wireless communication protocol is proposed where sensor node readings are transferred to a centralized station for monitoring and visualization. The sampling technique used was Non-probability purposive due to the fact that information gathered was important to contribute to better understand the problem by gathering information from the right personnel in the field of study. Through this approach, it is a perfect application of IoT and Wireless Sensor Network solution which has been achieved and provides advantages to the brewery industry. Moreover, the transfer of real-world processes to a digital world has been made possible which allows optimizing these processes.Through the project study, a simple and automated method with good accuracy was developed for estimating Brix and alcohol content during the fermentation process.
Characterizing Water Users through Frequent Patterns and Association Rules by Using Apriori Algorithm: A Case of Pangani Basin Tanzania
(IJST, 2024-12-07) Lyuba, Matimbila; Nyambo, Devotha; Sam, Anael; Tilahun, Seifu
Objectives: To identify the hidden patterns in the K-means clustered dataset for the Pangani Basin using the Apriori algorithm through frequent patterns and association rules to enrich cluster characteristics. Methods: Frequent patterns and association rule mining were used to discover the hidden attributes in the K-means clustered dataset. Measures of minimum support ranging from 0.5% to 5% and minimum confidence ranging from 50% to 100% were used to generate a manageable number of rules which were then filtered for redundancy. Lift value >1.0 was used to determine the rule’s interestingness while Arules and ArulesViz in R were used to visualize generated rules. Findings: Clusters one to four generated 25, 31, 47, and 49 rules respectively at a minimum confidence of 50% and minimum support of 2% in the first two clusters and 1% in other clusters. Furthermore, water users in cluster one were observed to abstract more water than the three clusters, while their water use fee also reflected on the amount they abstracted. In clusters two and three, water users identified the same amount of water source capacity but differed in the amount requested and water use fee. Water users in cluster four were identified with less water source capacity and fewer amounts abstracted than other clusters. However, their water use fee identified was higher than those in cluster three, with high water source capacity and high amount requested. Such a difference is attributed to the type of water use for cluster three users being domestically supplied through community water supply entities to help villagers access water. In contrast, the water use for users in cluster four is domestic and commercial. Novelty: When aggregated with the clustering observations, the identified association rules mining results provide a broad understanding of water users’ characteristics for better water allocation and rationing.
Data Synthesis Technique for Categorical Pestes Des Petits Ruminants (PPR) Data Using CTGAN Model
(Pre prints,org, 2023-05-11) Nyambo, Devotha; Mduma, Neema; Sinde, Ramadhani; Lyimo, Tumaini
Data scarcity is a significant challenge in the field of Machine Learning (ML), as data collection can be expensive, time‐consuming, and difficult, particularly in developing countries. This challenge is exaggerated on the need to use dataset for livestock disease predictions for early intervention and surveillance. To address this challenge, this paper presents a data synthesis method that has been used to accurately generate new data samples from few real‐world data. With much data available to train the ML models, overfitting is eliminated. We present the use of Generative Adversarial Networks mainly the Conditional Tabular Generative Adversarial Network to synthesize categorical data for training machine learning models for prediction of the Pestes des Petits Ruminants (PPR) disease. The results showed that training score became 0.89 and the cross‐ validation score was 0.87 after synthesized data was used with Random Forest algorithm. The resulting dataset can be used to support the prediction and surveillance of the Pestes des Petits Ruminants (PPR) disease. The proposed method can also be applied to any domain with categorical data, and has the potential to improve the performance of machine learning models with increased data availability.
Database Privacy: Design of User Privacy Preserving Central Bank Digital Currency: A Case of Tanzania
(Indian Society for Education and Environment (iSee), 2024-03-04) Minja, Godbless; Nyambo, Devotha; Sam, Anael
Objectives: This work aims to contribute towards Tanzanian Central Bank Digital Currency (CBDC) users’ privacy preservation. It proposes the design of a privacy preserving CBDC which might be issued by Tanzania’s Central Bank (CB), the Bank of Tanzania (BoT), which is currently in CBDC research phase. The work also aims to contribute to literature, the CBDC research being done by BoT, other CBs and CBDC stakeholders around the world. Methods: By using the Design Science Research (DSR) methodology, a privacy preserving CBDC design suitable for Tanzania was proposed, demonstrated and evaluated. This is the result of existing literature showing that different countries have different CBDC designs due to their differences in contexts and purposes for CBDC issuance. This consequently emphasized the fact that a CBDC design should not be treated as a one-size fits all solution. Findings: As opposed to the existing general and other country specific CBDC designs, we proposed a privacy preserving CBDC design suitable for Tanzania by consulting literature and taking into consideration the Tanzanian context. The design appears to be promising Tanzanian CBDC users’ privacy preservation though further work needs to be done. The work should not only be on practical evaluation of the proposed design but also on other factors impacting the success of CBDC projects. This will consequently further increase the success probability of CBDC projects, hence the potential for practical realization of CBDC project benefits. Novelty: Existing literature has shown that, considering the countries’ differences in context and CBDC issuance purposes, CBDC design should not be treated as a generic solution thereby obliging the need for country- specific CBDC designs. Consequently, the privacy preserving CBDC design suitable specifically for Tanzania consists of and provides an outline of privacy preserving interactions among the identified key Tanzanian CBDC participants or actors. The actors are the BoT, the intermediaries (i.e., other banks and payment service providers), Tanzania’s National Identification Authority (NIDA), financial transactions violation detection engine, and the expected CBDC users.
Dataset Development for Automated Grade Labelling of Virginia Flue-Cured Tobacco Leaves in Tanzania: A Focus on Stalk Leaf Position
(Indian Journal of Science and Technology, 2025-02-26) Nguleni, Faith; Nyambo, Devotha; Lisuma, Jacob; Kaijage, Shubi
Objectives: This study aimed at developing a dataset for automated grade labeling of Virginia flue-cured tobacco leaves based on stalk leaf position by focusing on quality, colour and anomalies. Methods: Virginia flue-cured tobacco leaves were collected from four Tanzanian tobacco regions: Tabora municipal, Uyui, Urambo and Kaliua. Canon 5D Mark III cameras with a Canon EF 100mm F/2.8L Macro IS USM lens were used to capture tobacco leaves. The collected data concentrated on the upper leaves of the tobacco plant, also known as leaf position. In Tanzania, the tobacco plant position is a very crucial entity during grade labeling processes. Findings: To fulfil the study’s intention, a dataset was created by collecting Virginia flue-cured tobacco leaf images. The study utilized our published dataset of Virginia flue-cured tobacco leaf images, which consisted of 49,779 high-resolution images with 22 grade labels (classes). Novelty: The important findings highlight the dataset's quality, making it crucial to develop automated systems for Virginia flue-cured tobacco leaf grade labeling processes based on stalk leaf position.labeling
Dataset of Virginia Flue-cured Tobacco Leaf images based on stalk leaf position for classification tasks: A case of Tanzania
(Elsevier, 2024-10) Nguleni, Faith; Nyambo, Devotha; Lisuma, Jacob; Kaijage, Shubi
Nicotiana tabacum is a kind of plant cultivated for its leaves used for manufacturing medicine and cigarettes. With the common name, the Tobacco plant is grown in many countries including China, Indonesia, Malawi and Tanzania just to mention a few. Literatures suggest a technical gap in the proper identification of grade labels for various parts of the plant. In addition, manual grading has resulted in various gaps and biases. To mitigate this, a data-driven grading solution is necessary. However, relevant datasets to train grade classifiers from various countries become of the essence. This article presents images concentrated on tobacco leaf plant position namely Leaf position which normally carries 23 grade labels. Due to high rainfall which swiped away the applied fertilizer on the tobacco plants in the farms, we failed to get images of one grade. Therefore, this research could capture and label 22 grade labels. Images of tobacco leaves based on the tobacco plant position were collected in Tanzania through participatory community research. Canon 5D mark III cameras with 100 mm micro lens were used to take pictures of tobacco leaves based on the tobacco plant position. Domain experts were used for image labelling and cleaning according to tobacco grade labels identified in Tanzania. The dataset carries 49,779 images, which can be used to develop machine learning models for tobacco leaf grade label identification. The collected dataset can be used to train models and enhance the performance of pre-trained models in any country of interest.
A Deep Learning Model for Classifying Black Sigatoka Disease in Banana Leaves Based on Infection Stages
(IJST, 2024-10-04) Nyambo, Devotha; Kambo, Edwin; Leo, Judith; Ally, Mussa
Objective: This research study aims to develop an efficient deep-learning model to detect and classify stages of Black Sigatoka disease in banana plants. Methods: In this study, deep learning techniques, specifically the basic Convolutional Neural Network (CNN) and VGG16 models, were used to address the challenge of identifying Black Sigatoka disease in banana leaves early on. The tests were conducted on a dataset containing labelled images of banana leaves, assessing their effectiveness based on criteria such as accuracy, precision, recall, and F1-score after adjusting hyperparameters for optimal outcomes. Findings: The results of the trials revealed that the basic CNN model attained a training accuracy of 96% and a validation accuracy of 89%, surpassing the performance of the VGG16 model. The VGG16 model, on the other hand, had a training accuracy of 92% and a validation accuracy of 89%. Across precision, recall, and F1 score measurements, the basic CNN model consistently outperformed the VGG16 model, with scores averaging 0.90 for all three metrics compared to VGG16’s precision of 0.80, recall of 0.75, and F1 score of 0.75. The CNN model demonstrated its efficiency by stopping training at 26 epochs, whereas VGG16 completed training in 21 epochs. This demonstrates its effectiveness in detecting Black Sigatoka while utilising minimal resources. Novelty: A significant component of this study is its emphasis on identifying the stages of Black Sigatoka disease, which is commonly overlooked in research. By studying disease progression, this study provides insights for early intervention and disease management, aiding efforts to lessen the impact of Black Sigatoka on banana farming.
A Deep Learning Model for Predicting Stock Prices in Tanzania
(Engineering, Technology & Applied Science Research, 2023-04-02) Joseph, Samuel; Mduma, Neema; Nyambo, Devotha
Stock price prediction models help traders to reduce investment risk and choose the most profitable stocks. Machine learning and deep learning techniques have been applied to develop various models. As there is a lack of literature on efforts to utilize such techniques to predict stock prices in Tanzania, this study attempted to fill this gap. This study selected active stocks from the Dar es Salaam Stock Exchange and developed LSTM and GRU deep learning models to predict the next-day closing prices. The results showed that LSTM had the highest prediction accuracy with an RMSE of 4.7524 and an MAE of 2.4377. This study also aimed to examine whether it is significant to account for the outstanding shares of each stock when developing a joint model for predicting the closing prices of multiple stocks. Experimental results with both models revealed that prediction accuracy improved significantly when the number of outstanding shares of each stock was taken into account. The LSTM model achieved an RMSE of 10.4734 when the outstanding shares were not taken into account and 4.7524 when they were taken into account, showing an improvement of 54.62%. However, GRU achieved an RMSE of 12.4583 when outstanding shares were not taken into account and 8.7162 when they were taken into account, showing an improvement of 30.04%. The best model was implemented in a web-based prototype to make it accessible to stockbrokers and investment advisors.
Design of the data-driven software application for identification, population monitoring, and risk assessment for lions in Serengeti Tanzania
(International Academy of Ecology and Environmental Sciences, 2024-09) Okey, Ambokile; Nyambo, Devotha; Kaijage, Shubi; Masenga, Emmanuel; Levi, Matana
This study presents a design of a Data-Driven software application for identification, population monitoring, and risk assessment for lions in Serengeti Tanzania. Lions’ populations have been declining due to poaching, overhunting, and other ecosystem factors resulting in unmet demands for tourism and ecological balance. Data-driven techniques can lower the negative consequences by providing mechanisms for lions’ management, risk assessment, and monitoring in selected wildlife reserves. Lion’s whisker spots, poaching rates, prey availability, human-conflict incidences, and pride size are key elements for achieving management, identification, monitoring, and risk assessment for lions. The software application design aimed at providing conceptual and logical requirements for the development of the application that will enhance lions’ monitoring and management efforts to protect their existence and contribution to the ecosystem. The study was conducted in the Serengeti ecosystem, including ecologists from the Tanzania Wildlife Research Institute Serengeti Wildlife Research Center, and information systems analysts. Through a mixed research methods approach, qualitative methods and incremental prototyping software development life cycle model were used to develop the specific requirements. Unified Modeling Language (UML) was used to model the requirements and led to the realization of design diagrams: application framework, database design, and artificial intelligence model workflows. The application should equip ecologists with tools to add and identify specific lions, monitor sightings, estimate population trends, assess risks for individual lions, and produce reports on monitoring and sightings. This design serves as a foundation for developing the data-driven software application for identification, population monitoring, and risk assessment for lions in Serengeti National Park Tanzania which will enhance monitoring and management activities of lions’ population non-invasively.
Developing an Agent-based Model for Evaluating the Effectiveness of Malaria Interventions in Nanyumbu and Masasi Districts, Tanzania
(creative common, 2024-11-28) Mfala, Celina; Nyambo, Devotha; Mwaiswelo, Richard; Mmbando, Bruno; Clemen, Thomas
Objective : Agent-based models and simulation (ABMS) can be utilized to understand the dynamism of transmission and the effect of interventions. We evaluated using ABMS, the efficacy of insecticide-treated bednets (ITNs) at different coverage levels and quality of houses for control of malaria in Masasi and Nanyumbu districts, Tanzania. Methods: The model was developed and simulated in Anylogic software with mosquitoes, humans, and the environment along with their attributes as agents. Using field data, buildings of different qualities were created to be human environment, and ITN use was assigned to respective human agents. Shapefiles were imported into the built-in global imaging system map in Anylogic for better placement of buildings using their coordinates, and coordinates of streams extracted from the study area map were used to allocate the aquatic environment of the mosquito agents. ITNs coverage scenarios of 16%, 40%, 64%, and 80% were simulated. The model was simulated for 90-day period and a model time-step was set to a day. The primary outcome was the prevalence of human agents with malaria infection at the end of the 90-day simulation period. Results: At the end of the 90-day simulation period and initial ITNs coverage of 16% (257/1607), the prevalence of malaria infection was 15.4% (248/1607). When the coverage was increased to 40%, 64%, and 80% malaria prevalence declined to 15.1% (242/1607), 14.1% (227/1607), and 13.9% (223/1607), respectively. ABMS clearly indicated that an increase in ITNs coverage was associated with a decline in the prevalence of infected humans and mosquito population in consistency with the field data. Novelty: This work is unique in a sense that it incorporated the data on house quality which has direct impact in malaria transmission.
Development of Railway Information System to Improve Railway Data Aggregation and Analysis in Tanzania
(Hindawi, 2023-05-11) Adam, Kitoi; Dida, Mussa; Nyambo, Devotha
For more than three decades, railway transportation in Tanzania has been in an on-and-off state even though a railway network exists. This is due to damaged tracks, a lack of proper management, and railway operational information. Recently, the Tanzanian government made efforts to revive railway transportation by reopening a few train routes and constructing a new and improved railway network. Even with revived operations, the digitalization process of railway data is still at a low pace as most data is populated in excel sheets for analysis; the major source of data being paper-based. With the use of a mixed research method, this paper provides an information system in the form of mobile and web applications, which provide a platform for populating railway data through the web application accessible to the railway corporation and disseminating railway information to the public through the mobile application. With these platforms, data aggregation and analysis have been made easier and more understandable than the use of excel sheets alone. The results show great possibilities for increased use of digital techniques such as web mapping, which contribute to higher data accuracy and better visualization of railway information that can be disseminated to the public.
External Services and their Integration as a Requirement in Developing a Mobile Framework to Support Farming as a Business via Benchmarking: The Case of NM-AIST
(TRANSACTIONS ON MACHINE LEARNING AND ARTIFICIAL INTELLIGENCE, 2020-10-30) Kyaruzi, John; Yonah, Zaipuna; Swai, Hulda; Nyambo, Devotha
Research on agricultural and rural development (ARD) systems in general, and farming as a business (FAAB) in particular, face the limitation of availability of credible and reliable benchmarking data, both for on-farm support for farm management decision making and off-farm support for research, investment and policy decision making. One of the main part of this limitation is to obtain reliable benchmarking data for decision making, both for current conditions and under scenarios of changed bio-physical and socioeconomic conditions. This paper presents a framework for mobile application development to support farming as a business via benchmarking (FAABB). This is done with a model that distinguishes between internal and external sources of data and between codified and computed information. Also, the paper demonstrates and emphasizes how integration should be considered as a requirement when developing a typical mobile application for ARD. The paper ends with a description of an ongoing research project at Nelson Mandela African Institution of Technology (NM-AIST) in Tanzania that aims to develop a new framework to facilitate development of mobile applications for FAABB.
Internet of things (IoT)‑based on real‑time and remote boiler fuel monitoring system: a case of Raha Beverages Company Limited, Arusha‑Tanzania
(Springer International Publishing, 2023-11-20) Ntambara, Boniface; Nyambo, Devotha; Solsbach, Andreas
RAHA Beverages Company (RABEC) is one of the banana wine production companies that utilizes fuel in steam produc- tion in Arusha-Tanzania, where fuel data conditions such as temperature, pressure, discharge, fuel level, and gas leakage with humidity were a challenge to monitor, which provoked boiler malfunction and plant breakdown. Today, RABEC manually uses a dropping stick into the fuel tank to monitor fuel data conditions which is time-consuming and gives inaccurate readings, inefficiency, fuel economy discrepancy, and accidents. This study aimed to design and develop an IoT-based fuel monitoring system. The flow meter, ultrasonic level, thermistor fuel temperature, humidity, and pressure sensors were used to gather fuel information where the GSM module was employed to send fuel data messages to the operator’s phone. An AT mega 328 microcontroller was used to process and analyse the fuel data and send them to the Thing Speak IoT platform using Wi-Fi connectivity. The results showed that when the fuel level was less than the thresh- old value, an operator was alerted by a refilling message via GSM technology. A pressure of 0.1psi, a fuel temperature of 120 ℃, and 80% of humidity, the system notifies the operator by an alert message to check the injector pressure and if the fuel–air mixture is perfect. In addition, these data were observed on the LCD and ThingSpeak webpage. To conclude, the developed system proved the best performance with a 99.98% of success rate with high accuracy, security, and efficiency rate compared to the current monitoring system.
IoT-based on boiler fuel monitoring system: A case of Raha Beverages company limited, Arusha-Tanzania
(Research Square, 2023-07-24) Ntambara, Boniface; Nyambo, Devotha; Solsbach, Andreas
RAHA Beverages Company (RABEC) is one of the banana wine production companies that utilize fuel in steam production in Arusha-Tanzania where fuel data conditions like temperature, pressure, discharge, fuel level, and gas leakage with humidity were a challenge to monitor them which provoked boiler malfunction and plant breakdown. Today, RABEC manually uses a dropping stick into the fuel tank to monitor fuel data conditions which is time consuming and gives inaccurate readings, inefficiency, fuel economy discrepancy, and accidents. This study aimed to design and develop an IoT-based fuel monitoring system. The flow meter, ultrasonic level, thermistor fuel temperature, humidity, and pressure sensors were used to gather fuel information where GSM module was employed to send fuel data messages to the operator’s phone. An AT mega 328 microcontroller was used to process and analyze the fuel data and send them to the Thing Speak IoT platform using Wi-Fi connectivity. The results showed that when the fuel level was less than the threshold value, an operator was alerted by a refilling message via GSM technology. At 0.1Psi pressure, fuel temperature of 120℃, and 80% humidity, the system notifies the operator by an alert message to check injector pressure and if the fuel-air mixture was perfect. In addition, these data were observed on LCD and ThingSpeak webpage. To conclude, the developed system proved the best performance with a 99.98% of success rate with high accuracy, security, and efficiency rate compared to the current monitoring system
Machine learning model for predicting Peste des Petits Ruminants
(IEEE, 2023-11-16) Nyambo, Devotha; Ngulumbi, Nguse; Mduma, Neema; Sinde, Ramadhani; Lyimo, Tumaini
Peste des petits ruminants (PPR) is a viral disease that affects small ruminants and is prevalent in many developing countries, particularly in Africa and Asia. It can spread through direct contact, air, and contaminated feed and water. PPR can result in significant economic losses and has a detrimental impact on small ruminant production and trade. Clinical signs include fever, respiratory distress, and diarrhoea, and prevention is primarily through vaccination with a live attenuated vaccine. In this study, 24 samples were selected, pre-processed and synthesized using the Conditional Tabular Generative Adversarial Networks (CTGAN) model. Feature extraction was performed, revealing difficult_breathing as the most important feature in predicting PPR in ruminants. The study used Random Forest Classifier which was fine-tuned using Bayesian Optimization to attain an accuracy of 91%.
Mobile Based Application for E-Services and E-Payments: a Study Case of Habari Node Public Limited Company in Arusha, Tanzania.
(EasyChair, 2023-03-07) Ritha, Umuhoza; Sam, Anael; Nyambo, Devotha
Technology is being involved in different sectors to improve service delivery. Habari Node PLC (Public Limited Company), located in Arusha, Tanzania, offers Internet Services and various additional ICT-based business solutions. The company has a website that is used to provide information related to the services they provide with their cost. However, the current website is not mobile user-friendly and is not integrated with an electronic payment to pay for those services because the fees are currently paid manually. This study aimed to develop a Mobile Based Application for E-services and E-payment which will allow the user to access all information related to the services provided by this company and be able to perform e-payment to the subscribes services. The payment will be made through mobile money or credit card, depending on the customer's choice.