Nowadays, people tend to spend considerable amounts of time reading text content on various electronic devices. The amount of perceived text often overwhelms us and we subconsciously use various mechanisms to process it as efficiently as possible. We can observe different kinds of reader behavior that depends either on the personality of the reader or on the characteristics of the text itself. Also, our choice of the reading style is heavily influenced by how interesting, important or comprehensible the given text seems to us.
In our work, we’ll analyze existing ways of determining reader literacy, text complexity and comprehension levels. Using gathered information, we’ll design an improved method for automatically identifying the subjective complexity of digital text. In order to get precise results, our method will rely on eye tracking data, user modeling and structural analysis of the given text. We’ll also evaluate the influence of the reader’s domain knowledge on the subjective complexity and understanding of the processed text. Finally, we’ll inspect some other possible factors such as text readability and its formatting.
Ultimately, the goal of our work is to compare the possible level of impact of all individual factors (mentioned above) on the subjective complexity of the digital text and its overall clarity to the reader.