Predicting the Algorithmic Time Complexity of Single Parametric Algorithms Using Multiclass Classification with Gradient Boosted Trees

• Published in: 2018 Eleventh International Conference on Contemporary Computing (IC3) pp. 1 - 6
• Main Authors: Sharma, Deepak Kumar, Vohra, Sumit, Gupta, Tarun, Goyal, Vipul
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.08.2018
• Subjects: Accuracy; Classification algorithms; Codes; Complexity theory; Machine learning; Machine learning algorithms; Manuals; Matplotlib; Pandas; Prediction algorithms; Pyplot; Python; Scikit; Tensorflow; Time complexity; xgboost
• ISBN: 1538668343; 9781538668344
• ISSN: 2572-6129
• DOI: 10.1109/IC3.2018.8530473

The amount of code written has increased significantly in recent years and it has become one of the major tasks to judge the time-complexities of these codes. Multi-Class classification using machine learning enables us to categorize these algorithms into classes with the help of machine learning tools like gradient boosted trees. It also increases the accuracy of predicting the asymptotic-time complexities of the algorithms, thereby considerably reducing the manual effort required to do this task, at the same time increasing the accuracies of prediction. A novel concept of predicting time complexity using gradient boosted trees in a supervised manner is introduced in this paper.