NWANKWO, CHUKWUMA MICHAEL and Covenant University, Theses (2022) A Predictive Model For Detecting Underage Voters using Deep Learning and Blockchain Technology. Masters thesis, COVENANT UNIVERSITY.
![[img]](http://eprints.covenantuniversity.edu.ng/style/images/fileicons/application_pdf.png) |
PDF
Download (372kB) |
Abstract
Elections around the world have become a major international concern since the inception of modern democracy. It is a fact that the success of any democracy depends largely on its electoral process. In conducting a free and credible election, the process must be transparent to be adjudged credible. The electioneering process begins with the compilation of a voter register; this register contains the details of every eligible voter as stipulated by law or guild lines that guilds the electoral process. As part of what makes up, the guidelines are age restrictions for every intended voter. It is forbidding by law in most countries for a child to register as a voter, but this is not so in reality in most countries, especially in a developing nation. Because the age restriction is not obeyed, this has resulted in the incidences of underage voters and disputed election outcomes. This work provides an efficient and effective solution for the above concerns, using multiple digital solutions. The model will be integrating a deep learning Convolutional Neural Network (CNN), an Interplanetary File System (IPFS), and an Ethereum Smart Contract Blockchain. The role of the CNN is to detect any underage individual who intends to register as a voter. The CNN is built with a pretrained dataset, and it was trained with an age classifier that grouped the age on the images on the data set into eight distinct groups. This age grouping will help the age predictive algorithm estimate and place every image on the camera in a unique age group. This will only produce a binary result, which is "eligible to voter or not eligible to vote." This outcome is based on the preset threshold cut-off on the age group. The Interplanetary File System (IPFS) will provide a large storage capacity that will allow for off-chain data storage and still provide the model with all the functionalities and benefits of the blockchain. It also provides a hashing function that will assign and identify every registered voter with a unique cryptographic Identity. This ID will prevent the storage of the same information into the database, in so doing, eliminating multiple voting. Finally, the blockchain will provide a voting platform where the model will be implemented. It will allow every registered voter with a unique ID to create an account and vote on the blockchain. The adapted CNN was tested and evaluated and shows 85.9% performance accuracy, and when compared against two other age predictive models, it recorded an increase of 1.2%. In comparison, the complete digital solution model recorded 95.3% in performance. We believe this model will perform even better when subjected to further research work.
| Item Type: | Thesis (Masters) |
|---|---|
| Uncontrolled Keywords: | Deep Learning, IPSF, CNN, Blockchain, Smart Contract, Age Estimation, E-Voting |
| Subjects: | Q Science > Q Science (General) Q Science > QA Mathematics > QA75 Electronic computers. Computer science Q Science > QA Mathematics > QA76 Computer software |
| Divisions: | Faculty of Engineering, Science and Mathematics > School of Electronics and Computer Science |
| Depositing User: | Mrs Hannah Akinwumi |
| Date Deposited: | 16 May 2022 17:50 |
| Last Modified: | 16 May 2022 17:50 |
| URI: | http://eprints.covenantuniversity.edu.ng/id/eprint/15840 |
Actions (login required)
 |
View Item |
