|
|
 |
| |
|
|
|
Enhanced Cultural Algorithm for Information Retrieval System |
|
|
|
PP: 1081-1094 |
|
|
|
doi:10.18576/amis/180514
|
|
|
|
Author(s) |
|
|
|
Doaa N. Mhawi,
Haider W. Oleiwi,
Ammar Aldallal,
|
|
|
|
Abstract |
|
|
| The number of web pages is rapidly increasing worldwide, making the information retrieval system (IRS) a crucial technology for search engines. Despite the enormous efforts made by researchers to improve the performance of IRSs, internet users still need to work on two main influential issues: imprecise content retrieved in response to their queries and accordingly resulting in irrelevant document retrieval. Thus, a high storage space was occupied while non- reasonable retrieval time was realized. In response to those challenges, this paper proposed a modified culture algorithm (MCA) allocated as an indexing method. The proposed system aimed to establish a way for indexing to retrieve specific documents rapidly, retrieve the relevant document for user queries/satisfaction, and reduce storage space. It applied the benchmark WebKb dataset with 8282 web pages that were semi-structured documents. According to the experimental findings, recall and precision metrics reached 99% for 40 test queries, whereas the storage size for document indexing occupied only 18 Megabytes. The relevant retrieved document was kept in a look-up table for quick access. Comparatively, the proposed system outperformed the state-of-the-art.
|
|
|
|
|
|
|
|
