The stock prices are considered as the reflection of fundamental information of the companies. In addition, the stock prices might be affected by events surrounding the company. The availability of a large amount of information has created the need for better knowledge discovery. However, most of that important information is in textual format, such as news articles, blogs, specialized magazines and other social media articles, which makes it changeling to automate the knowledge extraction. Textual information such as news articles, might be helpful to explain the performance of companies and markets, including events that can condition their behavior. Hence, it is considered as a rich source for analysis which can provide a better forecasting of future performance. In this paper, we propose a framework that attempts to determine the influence of events on stock market trends by extracting events patterns. Our work focuses on extracting those event patterns using an unsupervised method to retrieve types of word categories and semantic relationships. We tested our system on a large number of news articles from different sources across different industry categories to guarantee the data diversity in a certain stock markets. We conclude that the effectiveness of the proposed approach shows better performance against other approaches analyzed in this work.