A Self-Assessing Compilation Based Search Approach for Analytical Research and Data Retrieval

TL;DR

This paper introduces a self-assessing search algorithm designed to automate key components of meta-analysis, improving efficiency and relevance in large dataset research across various topics.

Contribution

The study proposes a novel, automated search approach that predicts feasibility and enhances source selection and relevance determination in meta-analytic research.

Findings

Average of 126 relevant sources retrieved per search

Efficiency of 19.55 sources per second

Potential for improved research methods across disciplines

Abstract

While meta-analytic research is performed, it becomes time-consuming to filter through the sheer amount of sources made available by individual databases and search engines and therefore degrades the specificity of source analysis. This study sought to predict the feasibility of a research-oriented searching algorithm across all topics and a search technique to combat flaws in dealing with large datasets by automating three key components of meta-analysis: a query-based search associated with the intended research topic, selecting given sources and determining their relevance to the original query, and extracting applicable information including excerpts and citations. The algorithm was evaluated using 5 key historical topics, and results were broken down into 4 categories: the total number of relevant sources retrieved, the efficiency given a particular search, the total time it takes to finish a complete cycle, and the quality of the extracted sources when compared to results from current searching methods. Although results differed through several searches, on average, the program collected a total of 126 sources per search with an average efficiency of 19.55 sources per second which, when compared and qualitatively evaluated for definitive results, indicates that an algorithm developed across all subject areas will make progress in future research methods.