Knowledge Acquisition & Hybrid Inference with a Stem-Based Approach (KISB)

The analysis of Arabic text is very important in works that require knowledge from the user. This work attempts to develop Arabic analyzer and then merge it with two modules; knowledge acquisition and hybrid inference. Each one of the two modules receives knowledge or part of knowledge (Arabic phrases) from many users. The knowledge acquisition system needs the Arabic analysis process to prevent: typing error and duplication. The hybrid inference system needs the Arabic analysis process: to prevent typing error, and to compute the probability of matching between the received premise and the stored one. In this work, the Arabic analyzer depends on stem-based approach. The results show that the Arabic analyzer is very effective in discovering: the wrong words and the similarity in meaning. To evaluate the proposed system, it was compared with another valued technique. And then, results of the comparison show that the proposed system have a lot of advantages over the traditionally shell technique.