| Popis: |
This paper presents a novel approach based on one of the signal processing tools in soft computing applied to Web information retrieval, namely wavelet transform. The influence of two parameters, wavelet functions (Mother wavelets) and decomposition level, on feature extraction and information retrieval ability of calibration model was investigated. The experimental results show that the proposed method performs accurate retrieval. This work is a step towards multilingual (English, Spanish, Arabic, Chinese (Simplified and Traditional), Korean, and Japanese) search engine. |
