|
|
 |
| |
|
|
|
Algorithm Comparison for Student-Supervisor Matching in Supervisorship System Development: K-Means vs. One-to-Many Gale-Shapley |
|
|
|
PP: 2417-2425 |
|
|
|
doi:10.18576/isl/121205
|
|
|
|
Author(s) |
|
|
|
Azamat Serek,
Meirambek Zhaparov,
|
|
|
|
Abstract |
|
|
| This paper presents a comparative analysis of the K-Means and Gale-Shapley algorithms for matching students to supervisors. The two algorithms are evaluated based on their performance in terms of preference satisfaction, balance of workload, time complexity, space complexity, and maximum and minimum workloads. The experimental results show that the Gale-Shapley algorithm outperforms the K-Means algorithm on all criteria, as shown in Table 4. Specifically, the Gale-Shapley algorithm achieves a preference satisfaction score of 0.74 and a balance of workload score of 0.5, compared to 0.34 and 0.2 for the K-Means algorithm. Additionally, the Gale-Shapley algorithm has a time complexity of O(num students × num supervisors) and a space complexity of O(num students + num supervisors), which is comparable to the K-Means algorithm. Finally, the Gale-Shapley algorithm has a maximum workload of 6 and a minimum workload of 3, compared to 15 and 3 for the K-Means algorithm. Based on the experimental results, the paper concludes that the Gale-Shapley algorithm is the superior algorithm for matching students to supervisors. It achieves a higher level of preference satisfaction and balance of workload than the K-Means algorithm, and it is still relatively efficient to run. The paper also discusses the advantages and disadvantages of both algorithms and provides recommendations for future work.
|
|
|
|
|
|
|
|
