Improve the Accuracy for Flight Ticket Prediction using XGBRegressor Optimizer in Comparison with Extra TreeRegressor Performance

• Published in: 2023 6th International Conference on Contemporary Computing and Informatics (IC3I) Vol. 6; pp. 2558 - 2562
• Main Authors: Saatwik Kumar, G.V., Jaisharma, K
• Format: Conference Proceeding
• Language: English
• Published: IEEE 14.09.2023
• Subjects: Extra Tree; Flight Ticket; Informatics; Machine Learning; Novel XGBRegressor Optimizer; Prediction algorithms; Random forests; Regression; Regression tree analysis; Software; Software algorithms; Training; Transport
• DOI: 10.1109/IC3I59117.2023.10397633

The important aspect of this study is to improve the accuracy for air transport Flight Ticket Prediction using Novel XGBRegressor Optimizer and Referred with Extra TreeRegressor Performance. We compare a Novel XGBRegressor Optimizer algorithm with the ExtraTree Regressor algorithm to enhance the system's accuracy. We use 40 sample sets for the comparison. The Novel XGBRegressor Optimizer algorithm is a new method that applies gradient boosting to optimize the parameters of the XGBRegressor model. The ExtraTree Regressor algorithm is a variant of the random forest algorithm that uses extremely randomized trees to reduce the variance of the predictions. Initialized Novel XGBRegressor is used for training the model. This helps to improve the execution of the system overall. Hence, the accuracy of the prediction is enhanced by the NXGBRegressor algorithm as it uses the best parameters during the prediction. ClinCalc software is used as aid to calculate the correctness of the setup under supervised learning. The alpha value is considered as 0.05, G-Power is 0.8, Confidential Interval (CI) of 95%. By performing the current research experiment, Novel XGBRegressor Optimizer has achieved an accuracy of 82.7 % , ExtraTree Regression has achieved an accuracy of 78.2%. The value of significance is found to be \mathrm{p}=0.000 after analyzing the results from the independent samples test and hence it is statistically significant as the value of \mathrm{p}=0.000 is found to be greater than (\mathrm{p} < 0.05) . In the present research article, the Novel XGBRegressor Optimizer is correlated with the ExtraTree algorithm. By succeeding the current research experiment it is found that the Novel XGBRegressor algorithm (82.7 % ) has more enhanced accuracy than the ExtraTree algorithm (78.2%).