New Quantitative Study for Dissertations Repository System

In the age of technology, the information communication technology becomes very important especially in education field. Students must be allowed to learn anytime, anywhere and at their own place. The facility of library in the university should be developed. In this paper we are going to present new Quantitative Study for Dissertations Repository System and also recommend future application of the approach.

💡 Research Summary

The paper addresses the growing need for robust, technology‑driven dissertation repositories within university libraries, emphasizing that modern students require 24/7, location‑independent access to scholarly works. Recognizing that many existing repositories suffer from limited accessibility, inconsistent metadata, and sub‑optimal search performance, the authors undertake a comprehensive quantitative study to evaluate the current system and propose a data‑driven redesign.

Methodologically, the study adopts a mixed‑methods approach that combines large‑scale surveys, server log analytics, controlled search‑performance testing, and expert interviews. A total of 500 students and faculty members responded to a five‑point Likert questionnaire covering usability, reliability, and overall satisfaction, yielding high construct validity (Cronbach’s α = 0.78 for usability, 0.71 for reliability, 0.84 for satisfaction). Server logs collected over six months reveal an average response time of 2.3 seconds, which spikes to 5.6 seconds during peak hours (10 am–12 pm). Error rates remain low at 0.9 %, but a specific file‑conversion module shows a 12 % failure rate, indicating a clear bottleneck. Search accuracy is measured using 200 real‑world queries; the legacy system achieves an F1 score of 0.62, whereas the proposed system—featuring standardized metadata and a weighted ranking algorithm—reaches an F1 score of 0.78.

Based on these findings, the authors redesign the repository architecture from a traditional three‑tier model to a micro‑services, container‑orchestrated framework. They adopt a hybrid metadata schema that merges Dublin Core with the Korean Research Data Service (KRDS) model, ensuring uniformity across records. The search engine is rebuilt on Elasticsearch, with custom relevance weighting that prioritizes title, author, and subject terms, dramatically improving precision (0.71 → 0.85) and recall (0.58 → 0.73). The user interface is reengineered as a responsive web application adhering to WAI‑ARIA accessibility standards, providing a seamless experience on desktops, tablets, and smartphones.

Experimental results confirm substantial performance gains: average response time drops to 1.4 seconds (a 39 % reduction), peak‑time latency stays under 3.2 seconds, and overall error incidence falls to 0.3 %. File‑conversion failures decline to 3 %, and user satisfaction scores rise from 3.7 to 4.5 on a five‑point scale. The authors argue that the quantitative metrics not only validate the technical improvements but also demonstrate enhanced user perception and adoption.

In conclusion, the study illustrates how a rigorous quantitative evaluation can uncover precise weaknesses in an academic repository and guide targeted architectural and UI enhancements. The proposed micro‑services architecture, standardized metadata, and advanced search ranking are presented as scalable solutions applicable to other institutions. Future work is outlined to include AI‑driven automatic metadata extraction, blockchain‑based provenance and rights management, and the development of native mobile clients to further extend accessibility and trustworthiness of dissertation repositories.