Integration and analysis of unstructured data towards database optimization and decision making using deep learning techniques
| dc.contributor.author | Maxima, Ainomugisha | |
| dc.date.accessioned | 2024-10-29T08:07:33Z | |
| dc.date.available | 2024-10-29T08:07:33Z | |
| dc.date.issued | 2024-06 | |
| dc.description | A thesis submitted to the school of mathematics and computing in partial fulfilment of the requirement for the award of a degree of Master of Science in software engineering of Kampala international university | |
| dc.description.abstract | This thesis addresses the challenge of integrating unstructured data into a Relational Database Management System (RDBMS). The increasing volume and variety of unstructured data pose significant challenges for organizations seeking to leverage such data for decision-making. Traditional RDBMS are not well-equipped to handle unstructured data due to their structured nature, leading to inefficiencies in data storage and analysis. To overcome these challenges, a model is developed to automatically integrate unstructured data into a Relational Database Management System (RDBMS). The objectives include designing a classification model, implementing it for data integration and analysis, optimizing it for database optimization and decision support, and validating its effectiveness. The model efficiently extracts relevant information from categorized unstructured documents, facilitating structured database construction. The study rigorously followed a data science research methodology, encompassing data collection, model development, implementation, testing, evaluation, and validation. Results show significant performance improvement with the incorporation of LSTM layers, notably achieving an accuracy boost from 83.2% to 94.6% in receipt image processing. Similar improvements were observed across precision, recall, and F1-Score metrics. This accomplishment substantially addressed the hurdles associated with processing and analyzing unstructured data. In conclusion, the researcher strongly recommends the adoption of this model for the analysis of unstructured data. Future research could focus on further optimizing the model's performance and scalability, exploring additional deep learning techniques, and extending its applicability to other domains. | |
| dc.identifier.uri | http://hdl.handle.net/20.500.12493/14548 | |
| dc.language.iso | en | |
| dc.publisher | Kampala International University | |
| dc.title | Integration and analysis of unstructured data towards database optimization and decision making using deep learning techniques | |
| dc.type | Thesis |