Preprocessed C# Source Codes for Machine Learning

<p>The dataset comes from the HackerRank site, 329,937 C# source codes of 22 tasks were collected and all verified by unit tests.</p> <p>During the download process, source codes received only a unique serial number instead of the user name who solved the task and stored inside the &#39;task_name/origin&#39; folder. After collecting the data, a new database was created, which included cleaned-up versions of the source codes (&#39;task_name/cleaned&#39; folders contains). Finally, a third set of data was extracted from this cleaned-up version, where a delimiter was inserted before and after each elementary expression to support easy processing and analysis processes (&#39;task_name/reduced&#39; folders contains). Inside the &#39;task_name&#39; folder three csv files, which contain the equality checking result. The compressed folder also contains a vector space (and related files) made from the reduced data set. These four files are directly in the main folder.</p>