Abstract
This paper analyses challenges, and provides a model and a framework for mining Arabic tweets to measure customer satisfaction toward telecom companies in Saudi Arabia, to predict the ratio of customer churn and overcome the specific challenges with the semantic sentiment analysis of Arabic text.