Arabic Information Retrieval

Effectiveness of alternate text processing techniques for Arabic retrieval are investigated in this study. The techniques that were studied are term weighting schemes, stemming, and stop words elimination. This research explored the effect of different weighting schemes on the retrieval effectiveness in Arabic Information Retrieval. The weighting schemes that were examined are the inverse document frequency weight, probabilistic weighting, and statistical language modeling. With these weighting schemes three stemming algorithms for Arabic text were used and three stoplists were created in order to combine the statistical approaches with linguistic approaches to reach an optimal performance. The data set that was used in the experiment is the LDC (Linguistic Data Consortium) Arabic Newswire data set.