http://repository.aust.edu.ng/xmlui/handle/123456789/374 This collection contains Computer Science Student's Theses from 2009-2024 2026-06-09T19:02:41Z http://repository.aust.edu.ng/xmlui/handle/123456789/5206 Fast and Accurate Feature-based Region Identification Maduakor, Francis There have been several improvements in object detection and semantic segmentation results in recent years. Baseline systems that drive these advances are Fast/Faster R-CNN, Fully Convolutional Network and recently Mask R-CNN and its variant that has a weight transfer function. Mask R-CNN is the state-of-art. This research extends the application of the state-of-art in object detection and semantic segmentation in drone-based datasets. Existing drone datasets was used to learn semantic segmentation on drone images using Mask R-CNN. This work is the result of my own activity. I have neither given nor received unauthorized assistance on this work. 2019-06-20T00:00:00Z http://repository.aust.edu.ng/xmlui/handle/123456789/5205 Multiple String-Matching Using Wavelet Matrix and Burrows-Wheeler Transform (Bwt) Adam, Saleh Adam The problem of multiple string matching is fundamental in computer science, with applications in bioinformatics, text mining, and information retrieval. Traditional methods struggle with large datasets due to high computational and memory requirements. This research proposes a novel algorithm that combines the Burrows-Wheeler Transform (BWT) for text compression and the Wavelet Matrix (WM) for efficient pattern search. The proposed method achieves faster search times, lower memory usage, and effective compression, particularly for repetitive datasets like DNA sequences. Experimental results demonstrate that the method performs better compared to existing algorithms. This work contributes to the advancement of efficient and scalable multiple string-matching techniques, with potential applications in large-scale text processing and bioinformatics. Keywords: Algorithms, Text Compression, Wavelet Matrix, Burrows-Wheeler transform, Multiple String matching 2024-02-18T00:00:00Z http://repository.aust.edu.ng/xmlui/handle/123456789/5135 How Machine Learning Can Evaluate the Influence of Socioeconomic and Climatic Factors in Agricultural Yield: A Case of Nigeria Dappa, Tamuno-Opubo Godwin The major international agencies in charge of nutrition are becoming increasingly concerned about global agricultural production in particular. Food insecurity has emerged in some populated areas, including Africa, as a result of the increased worldwide need for food as a result of record population growth. Climate change and its variability are two additional factors that contribute to world food insecurity. Furthermore, agricultural policy officials, farmers, and decision-makers require advanced technologies in order to make timely strategies or policies that will have an effect on the quality of crop harvests. Machine learning and other new, powerful analytical techniques made possible by big data technologies have already proven useful in a number of industries, including biology, finance, and medicine. The yield of three major crops, including cocoa, sesame, and cashew, at the national level in Nigeria during the course of the years spanning 1990 to 2020 is forecasted in this study using a machine learning-based prediction method. We used climatic, agricultural yield, and socioeconomic data to help policymakers and farmers anticipate the yearly agricultural outputin Nigeria. We employed k-nearest neighbors, a decision tree, and random forest. We also employed a hyper-parameter tweaking technique through cross-validation to enhance the model and avoid overfitting. For sesame, the accuracy of the Decision Tree model was the highest, having a test accuracy of 97.92% for socioeconomic and climatic factors combined, while the KNN model did the best with a test accuracy of 99.71% for climatic components separately. The accuracy of the Random Forest model was 87.54% for climatic elements alone and 87.64% for socioeconomic and economic factors together. For cocoa, the Decision Tree model had an accuracy of 89.49% for socioeconomic and climatic factors combined and 89.51% for climatic components alone, while the KNN model had the best accuracy of 90.71% for climatic elements alone. For socioeconomic and climatic factors taken together, the Random Forest model's accuracy was 87.82%; for climatic components alone, it was 88.83%. For cashew nuts, the accuracy of the KNN model was 78.38% for socioeconomic and climatic components combined and 99.81% for climatic factors alone, compared to 88.27% for socioeconomic and climatic elements combined and 86.58% for climatic factors alone for the Decision Tree model. For both socioeconomic and climatic components combined, the Random Forest model's accuracy was 98.50%, while for climatic factors alone, it was 98.75%. In conclusion, the Random Forest model outperformed the KNN and Decision Tree models across all crop and factor combinations. Our findings indicate that machine learning algorithms can be used to forecast crop yields with reasonable accuracy when socioeconomic and meteorological variables are combined Main Thesis 2023-05-10T00:00:00Z http://repository.aust.edu.ng/xmlui/handle/123456789/5123 Neural Collaborative Filtering and Autoencoder Enabled Deep Learning Models for Recommender Systems Arnold, Kwofie Finding important and useful information is getting harder as much more information is available online. The challenge for content producers is to deliver the appropriate content to the appropriate consumers while making it challenging for users to access that content. The foundation for overcoming these difficulties is provided by recommender systems. Traditional methods like Collaborative Filtering (CF) and Content-Based Recommender Systems have historically been successful in this field of study but are now challenged by problems with data sparsity, cold start, and non-linearity interaction. Evidently, several academic areas, like image detection and natural language processing (El-Bakry, 2008), have shown great interest in deep learning due to outstanding performance and the alluring quality of learning intricate representations. The impact of deep learning is recently showing good advancement when applied to recommender systems research (He, 2008). In this research We dive deep into the Autoencoder and Neural Collaborative Filtering based deep learning models and their implementation on classical collaborative filtering. The research also evaluates the performance of both models and outlines loopholes which can further be improved in future works. Main Thesis 2022-11-15T00:00:00Z