Measuring Text Difficulty

Let’s say that you work for a company that is developing a language-learning program. You are asked to come up with an algorithm that can measure the difficulty of reading some written text for non-fluent speakers of a language.

For example - we want to be able to quantify if a sentence is difficult to comprehend (⁹⁄₁₀ fluency) or easy to start out with (¹⁄₁₀ fluency).

How would you go about building this algorithm?