Summary
Ph.D. from UTM Malaysia. The total Impact Factor is 51.901 for research. Teaching experience of 21 years and remained HoD for 2 tenures at COMSATS University Islamabad
Academic Background
PhD
( Testing Profile for Software Reliability Engineering Models)
University Technology Malaysia
2017
MS
(RC5 Text encryption image embedding)
UET Taxila
2008
MCS
( Software Engineering)
Bahria University
2002
Experience
Assistant Professor
COMSATS University Islamabad
17-Jan-2010 - 13-Mar-2023
Lecturer
Air University
01-Jun-2004 - 16-Jan-2010
Teaching Assistant
Bahria University
01-Feb-2002 - 31-May-2004
Honours and Awards
Best HoD
COMSATS University Islamabad
01-Jan-2022
Best Researcher
MySec Malaysia
01-Apr-2015
Journal Publications
Principal Component Analysis (PCA) and feature importance-based dimension reduction for Reference Evapotranspiration (ET
) predictions of Taif, Saudi Arabia
02-Jul-2024
Reference Evapotranspiration (ET
) is fundamental to irrigation water management but challenging to calculate due to requirements of many weather parameters for standard Penman–Monteith (PM) method of Reference Evapotranspiration (ET
) calculation. Many machine-learning approaches were proposed for the simplification of Reference Evapotranspiration (ET
) predictions. There is also a need to explore the possibilities of daily Reference Evapotranspiration (ET
) predictions for the desert climate of Taif, Saudi Arabia. The study proposed machine learning-based daily Reference Evapotranspiration (ET
) predictions of Taif, Saudi Arabia. The weather data of Taif, from 2001 to 2023 is used to train and evaluate the performance of the Decision Tree Regressor (DTR), Extreme Gradient Boosting Regressor (XGBoostR), Random Forest Regressor (RFR), and Light Gradient Boosting Machine Regressor (LightGBMR) based machine learning models. The LightGBMR model outperformed other models for daily Reference Evapotranspiration (ET
) predictions of Taif, with a coefficient of determination (R
) of 0.998, a Mean Squared Error (MSE) of 0.016 mm day−1, a Root Mean Squared Error (RMSE) of 0.128 mm day−1, and Mean Absolute Error (MAE) of 0.093 mm day−1, using twelve weather parameters. The feature importance of the LightGBMR model shows that weather parameters for ET
predictions of Taif are important in order of wind speed (u
), maximum temperature (T
), relative humidity (R
), solar radiation (R
), extraterrestrial radiations (R
), saturation vapor pressure (
), actual vapor pressure (e
), minimum temperature (T
), net long-wave radiation (R
), number of possible sunshine hours (N), net radiations (R
), and number of actual sunshine hours (n). The Principal Component Analysis (PCA) shows that the top five weather parameters can capture 99.6% variance for ET
predictions of Taif. The LightGBMR model trained with top five weather parameters performed equally well as the LightGBMR model trained with all weather parameters, with a R
of 0.998, a MSE of 0.016 mm day−1, a RMSE of 0.128 mm day−1, and MAE of 0.094 mm day−1. The study provides valuable insights for using appropriate weather parameters for the Reference Evapotranspiration (ET
) predictions of Taif, Saudi Arabia.
Conference Publications
Courses
- Software Reliability Engineering
- Software Quality Engineering
- Advanced Computer Architecture
- Research Methodology
- Advance Operating Systems
- HCI
IoT based Smart Agri Solution
SRGP
Details
Namal University