In this project, relation extraction in Turkish text will be implemented using pattern matching and supervised machine learning techniques. First, the existing literature will be reviewed. Then, a corpus will be collected and the relation types that will be extracted will be identified. This will be followed by developing different types of algorithms for extracting the relations.
In this project, a Turkish-English parallel corpus will be created. We will start by analyzing an already existing corpus. Then, new data will be collected and sentences will be aligned using alignment algorithms. When the corpus is ready, machine translation applications between Turkish and English will be implemented using some existing tools. We will develop a statistical translation and a neural translation system.