A5103112270- Imbalanced Data Classification Techniques
- Machine Learning and Data Classification
- Data Mining Algorithms and Applications
- Human Pose and Action Recognition
- Software System Performance and Reliability
- Cognitive Radio Networks and Spectrum Sensing
- Robot Manipulation and Learning
- Online Learning and Analytics
- Reinforcement Learning in Robotics
- Advanced MIMO Systems Optimization
- Radar Systems and Signal Processing
- Microwave Imaging and Scattering Analysis
- Advanced SAR Imaging Techniques
- Water Quality and Pollution Assessment
- Text and Document Classification Technologies
- Adversarial Robustness in Machine Learning
- Wireless Communication Networks Research
The Future University
2022
Eastern Mediterranean University
2016-2018
University of Bahri
2016-2018
During these decades, data mining has become one of the effective tools for analysis and knowledge management system, so that there are many areas which adapted approach to solve their problems. Using in education enhance system is still relatively new. This paper focuses on predicting instructor performance investigates factors affect students’ achievements improve quality. Turkey Student Evaluation records dataset considered run different classifier such as J48 Decision Tree, Multilayer...
Pruning is applied in order to combat over-fitting problem where the tree pruned back with goal of identifying decision lowest error rate on previously unobserved instances, breaking ties favour smaller trees high accuracy. In this paper, pruning Bayes minimum risk introduced for estimating risk-rate. This method proceeds a bottom-up fashion converting parent node subtree leaf if estimated risk-rate that less than risk-rates its leaf. paper proposes post-pruning considers various evaluation...
This paper introduces v0.5 of the AI Safety Benchmark, which has been created by MLCommons Working Group. The Benchmark designed to assess safety risks systems that use chat-tuned language models. We introduce a principled approach specifying and constructing benchmark, for covers only single case (an adult chatting general-purpose assistant in English), limited set personas (i.e., typical users, malicious vulnerable users). new taxonomy 13 hazard categories, 7 have tests benchmark. plan...
Along with the enormous development of computer systems and fast spread internet, data processing analysis have become a significant concern. Different soft computing techniques been introduced to extract valuable information from data. These applied in different areas reflected useful promising results. In this paper, novel decision tree algorithm combined linear regression is proposed solve classification problem. The method Turkey Student Evaluation Zoo datasets that are taken UCI Machine...
The Forward scattering radars (FSRs) are special types of Bistatic in which detected targets should exist the narrow baseline to obtain their tracking at an angle 180 degree. This gives radar several features such as target classification makes FSR more privileged comparison traditional systems. Existing research works concerning ground detection and have utilized neural network for identification processes compared it other statistical models terms signal complexity. However, these...
This study is the first to be conducted in Sharurah governorate, Saudi Arabia, on groundwater. In this study, correlation analysis and water quality index WQI were used analyze thirty data points of groundwater some fields (15wells) city its outskirts. The utilized coefficient relationship between different physicochemical parameters such as Hydrogen ion concentration (pH), electrical conductivity (EC), turbidity (Turb), total dissolved solids (TDS), hardness (TH), chloride (Cl-), sulfate...
In imitation and reinforcement learning, the cost of human supervision limits amount data that robots can be trained on. An aspirational goal is to construct self-improving robots: learn improve on their own, from autonomous interaction with minimal or oversight. Such could collect train much larger datasets, thus more robust performant policies. While learning offers a framework for such via trial-and-error, practical realizations end up requiring extensive reward function design repeated...
In this paper, a post-pruning method known as zero-one loss function pruning (ZOLFP) that is based on introduced. The proposed ZOLFP minimizes the expected loss, rather than evaluating misclassification error rate of node and its subtree. subtree pruned when less or equal to sum leaves. experimental results demonstrate outperforms. Un-pruned C4.5 Decision Tree (UDT-C4.5) algorithm, reduced (REP), minimum (MEP) in terms performance accuracy all used datasets. It also shown complexity not more...