Assessing the Accuracy Level of University-Based Website-Based Search Engines Using F-Measure and Hellinger

Sep 10, 2026

Websites serve as essential information media in university environments, especially for learning-related activities such as library references, administrative updates, and important announcements. When universities rely on social media platforms for communication, accuracy and relevance often decline due to promotions, hoax content, and unrelated information, making the overall information flow excessive. This work optimizes campus website search engines to provide targeted and accurate results, evaluating accuracy levels and the impact of retrieval outcomes. Measurement uses quantitative F-Measure and the Hellinger method.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages12

Reading Time30 min

AccessFree to download

Introduction

Background and Motivation

Prior Research

Search Engine Development for Document Similarity

Measurement Approach

Assessing the Accuracy Level of University-Based Website-Based Search Engines Using F-Measure and Hellinger Irfan Santiko1,*, Gerry Andriana2 1Information System Departement, Amikom Purwokerto University, Banyumas, Indonesia 2Informatics Departement, Amikom Purwokerto University, Banyumas, Indonesia [Email:](Email:1irfan.santiko@amikompurwokerto.ac.id)[1](Email:1irfan.santiko@amikompurwokerto.ac.id)[irfan.santiko@amikompurwokerto.ac.id](Email:1irfan.santiko@amikompurwokerto.ac.id), [2](2gerry_an@gmail.com)[gerry_an@gmail.com](2gerry_an@gmail.com) Abstract Websites are an information medium that is becoming something that is needed in this era. Including the media within the campus environment. The problem is that the campus as a forum or place for student learning is considered less than optimal in presenting information on student learning activities. For example, library reference information, administration, important announcements, and other similar information. The current solution is that universities use social media platform communication media which are considered accurate, which actually adds to problems when the media is used not in accordance with its function, such as promotions, hoax information and irrelevant information. This causes the information to become too massive so that the level of accuracy and relevance is reduced. The author's solution is to optimize the search engine on the campus website platform to be used as an absolute information medium. So the information obtained will be more targeted and accurate. Starting from measuring the level of accuracy to the impact of the results will be discussed in this article. The technique used to measure accuracy is a quantitative technique consisting of the F-Measure and the Hellinger Method. As a result, the campus will know that to distribute related news, the campus can find out keywords that are considered strategic in every report on the media website. Keywords: Search Engine, Website, Keyword Measuring, News Accuracy 1. INTRODUCTION A similarity detection tool that can show the degree of variance between text documents was developed for Indonesian text documents in an effort to reduce plagiarism [1] . The document similarity detection model uses the Hellinger algorithm technique in an effort to achieve high accuracy [2], [3] . Hellinger's precalculation method finished the tokenization, filtering, and word root construction processes. Word root formation is done using the Indonesian rule basis stemmer approach, whereas stop-word tuning is employed for filtering [4],[5]. A perception test is performed utilizing an extensive Indonesian lexicon to determine the relevance and irrelevantness of the terms being searched, prior to evaluating the similarity identification technique. The results of testing the document similarity finding tool with the Hellinger technique [6] . There are many words that have been put together in a sentence in a document that shares some similarities with other documents, or with many other documents that already exist. The search results with visible results show how similar the documents are to one another. Search engines operate in accordance with their purpose, which is to give users access to pertinent information [7], [8] . Due to the employment of various algorithms, search engines generate information with varying outcomes. Textual documentation can be particularly helpful if it is easily accessed and handled within an integrated database. There are not many actors or datainterested parties that have implemented integrated documents [9] . Prefixes, inserts, and suffixes are what make Indonesian unique as a text document object that may be utilized for a variety of purposes, including data processing. Periodically, the volume of data grows; if it is not adequately handled, it will not be able to be utilized to its fullest extent. The same material in data will keep repeating and piling up. As a resu

FAQs

Why do universities need a website-based search engine instead of relying on social media?

Social media communication can include promotions, hoax information, and irrelevant content, which makes information too massive and reduces accuracy and relevance.

What methods are used to measure the accuracy level of the search engine results?

Accuracy is measured using quantitative techniques, specifically F-Measure and the Hellinger method.

What is the main objective of this research?

To optimize campus website search engine usage as an accurate information medium and to discuss accuracy measurement and the impact of the retrieval results.

Sep 10, 2026

Uploaded by Felix Montgomery