Modeling textual document classification

We investigate existing rule-based techniques for automatic textual document classification. The weakness of these techniques are identified. We propose a new technique known as the IBRI algorithm by unifying the strengths of rule-based and instance-based methods. Our algorithm adapts to the characteristic of text classification problems. Some experiments have been conducted to demonstrate the effectiveness of our IBRI algorithm. Moreover, we compare the performance with an existing rule-based and instance-based algorithms. The results show that our IBRI performs better most of the time