xap/SelfCode

In code comprehension tasks, students may be asked to explain critical steps of the code. The ability to automatically assess these self-explanations offers a unique opportunity to understand the current state of student knowledge, recognize possible misconceptions, and provide feedback. Annotated datasets are needed to train Artificial Intelligence/Machine Learning approaches for the automated assessment of student explanations. To answer this need, we present a novel corpus called SelfCode which consists of 1,770 sentence pairs of student and expert self-explanations of JAVA code examples along with semantic similarity judgments provided by experts.

Data Set Example

Line of Code	Crowd Sourced Explanation	ExpertExplanation	Annotation Score
int [] arr = { 1, 2, 3};	Declares the array we want to use for our assignment	We initialize the array of type int to hold the specified numbers.	4
int num = 15;	variable decleration: declares the number we are trying to divide	We could initialize it to any positive integer greater than 1.	1
divisor += 1;	if the loop condition is true, then we increment the divisor by 1	When the divisor is not a factor of the number, we increment the variable divisor by 1.	3
int seconds = scan.nextInt();	Get the number entered by the user.	We read the seconds by calling the nextInt() method because this input is an integer.	2
System.out.println("Enter an integer: ");	Ask the user to enter an integer.	We prompt the user to enter an integer.	5

Data Set Stats

Annotation Label	No. of sentence pairs
1	529
2	507
3	419
4	253
5	62

Cite Our Work

If you use the SelfCode data set, please cite our publication:

@article{Chapagain_Risha_Banjade_Oli_Tamang_Brusilovsky_RUs,
    title = {SelfCode: An Annotated Corpus and a Model for Automated Assessment of Self-Explanation During Source Code Comprehension},
    volume = {36},
    url = {https://journals.flvc.org/FLAIRS/article/view/133385},
    number = {1},
    journal = {The International FLAIRS Conference Proceedings},
    author = {Chapagain, Jeevan and Risha, Zak and Banjade, Rabin and Oli, Priti and Tamang, Lasang and Brusilovsky, Peter and Rus, Vasile}
}