Journal of Computational Science and Data Analytics

Eucalyptus Trees Nutrition Deficiency Detection Using Deep Learning Techniques

2026-02-24T12:39:25+00:00

Eucalyptus is one of the most widely cultivated tree species in Ethiopia and as a major source of fuel wood and construction material. A few researchers have attempted to do their research in other countries try to solve Eucalyptus tree nutrient deficiency by identifying the leaf part of the tree. According to this idea many nutrition deficiencies that affect Eucalyptus tree vegetation growth in Ethiopia. In this article, a model was developed using the deep-learning techniques, that developing Eucalypts trees nutrition deficiency detection. The study employs four distinct architectures, using one of a cloud software tool that known as "Google colab" or "Google collaborator"; MobileNet, ResNet50, DenseNet, and InceptionV3 to analyze and classify nutrient shortages from Eucalyptus tree leaf images. We collected 843 images, from its dataset 80% of the data are utilized for the training set, or a total of 674 images for all class; the remaining 10% are used to test the model and 10% images are used for the validation set for each class, MobileNet received a more hopeful result scored based on the report of classification parameters between all models. The Eucalyptus nutritional deficiency was correctly identified based on the collected images. A thorough investigation of the above mentioned models during the process of differentiating between the nutrient-deficient and healthy managed the following nutrients' Boron (B), Calcium (Ca), Iron (Fe), Magnesium (Mg), Nitrogen (N), Phosphorus (P), Potassium (K), and Healthy on images of the leaves. Additionally, the model's data performance could be improved. Better forest management will result from this research's difficult way of monitoring Eucalyptus tree health, which will benefit precision agriculture production.

Artificial Intelligence and Social Media Data as an Alternative Source of Insights During Pandemics: A Case Study of Twitter for COVID-19 in Tanzania

2025-09-18T17:05:21+00:00

Recently, social media has become one of the major sources of information for different sectors, including healthcare. During COVID-19, there was a lot of community engagement on social media than traditional physical engagements. Tanzania, in particular, had a different approach to tackling the COVID-19 pandemic as the country didn’t really practice lockdown, and very little information was shared by the authorities in traditional media houses. To understand what really happened during the pandemic, we tried to investigate social media as people were sharing information rather than the traditional media houses, and find the correlation with other sources. In this study, we extracted and analysed four-day periods of Twitter posts of Kinondoni district, Dar es Salaam, Tanzania. The four days were picked intentionally as it was the time when Tanzania started approaching the pandemic, as the rest of the world changed its course. Most of our analysis results significantly correlate with the results reported by the government during the same period, 21st - 24th April 2020. We further performed an analysis of how much the COVID issue was discussed online. As the WHO and many governments around the world have been providing education to people on how to protect themselves and slow down the spread, none have had a way to measure how well people were educated. We analysed 20,421 tweets of Kinondoni district, the most populous district in Dar es Salaam, and where many expats live, and found out how well people were educated about the CORONAVIRUS disease. We further created an Artificial Intelligence algorithm, a Deep learning to be specific, which has been able to classify tweets into COVID and Non-COVID classes with an accuracy of 93%. Our results mean that social media data analysis can be used as a tool for topic modelling to detect the most trending topics like disasters, election events, and epidemics.

Ethiopian Original, Different Side and Counterfeiting Bank Notes of Both Side Verification Using Deep Learning Approach

2025-12-11T08:48:23+00:00

Counterfeit currency is a growing concern in Ethiopia, affecting economic stability and public trust. This study presents a deep-learning system that uses Convolutional Neural Networks (CNNs) to detect counterfeit Ethiopian banknotes of 100-birr and 200-birr denominations. Both front and back sides of genuine and counterfeit samples are analyzed. A dataset of 1,500 images was augmented to 7,800 and split into eight classes. Three pre-trained CNNs—MobileNet, DenseNet121 and InceptionV3—were fine-tuned and compared with a custom CNN. MobileNet achieved the highest accuracy (96.58%), followed by DenseNet121 (95.17%) and InceptionV3 (93.99%). The custom CNN reached 94.20%. The results demonstrate the suitability of transfer learning for scalable and accurate currency authentication in banking and commercial applications.