Amira M. Idrees

Enhancing Iterative Dichotomiser 3 algorithm for classification decision tree

Data mining tasks such as clustering and classification have proved to highly impact various fields such as business, including the banking sector, as well as medicine, including the radiology sector. As the decision‐making process is critically dependent on the availability of high‐quality information presented in a timely and easily understood manner, the successful application of efficient data mining approaches is a great support for achieving the required target in the available time. This study presents an enhancement for the Iterative Dichotomiser 3 (ID3) classification decision tree algorithm based on two related approaches, namely, data partitioning and parallelism. The study applied the proposed algorithm in the banking and radiology sectors; as data have been classified to the defined fields’ clusters, the processing time and the results’ accuracy parameters have been compared with the ID3 algorithm and have proved an enhancement in both parameters. WIREs Data Mining Knowl Discov 2016, 6:70–79. doi: 10.1002/widm.1177

A healthy nutrition expert system for children

Healthy food is one of the most critical aspects for the family, especially for their children as healthy diet for children results in better life including the increase of their ability to learn, exercise, and even behaving. Therefore, a need for a full support for providing the children with healthy food is a vital target to reach. In this paper, we propose a prototype for a children nutrition expert system which aim is to provide its users with the nutrition expertise. It generates healthy meals for children in different ages according to different criteria including their growth stage, gender, and their health status. A web application is developed and two case studies are applied to demonstrate how the proposed prototype can be applied for determining childs nutrition.

A proposed configurable approach for recommendation systems via data mining techniques

ABSTRACT This study presents a configurable approach for recommendations which determines the suitable recommendation method for each field based on the characteristics of its data, the method includes determining the suitable technique for selecting a representative sample of the provided data. Then selecting the suitable feature weighting measure to provide a correct weight for each feature based on its effect on the recommendations. Finally, selecting the suitable algorithm to provide the required recommendations. The proposed configurable approach could be applied on different domains. The experiments have revealed that the approach is able to provide recommendations with only 0.89 error rate percentage.

Adaptive Classification Method Based on Data Decomposition

Knowledge discovery is one of the vital fields which strongly supports decision making by applying different techniques based on the targeted field and the required information. Focusing on clustering and classification techniques, this paper presents an approach for adapting one of the classification algorithms for supporting decision making procedure in radiology data analysis field. The proposed adaptation is based on dividing the analysis problem by data partitioning and individually examining against each cluster, with applying the classification algorithm in a parallel approach. The proposed approach has proved to produce higher results accuracy with minimization of time when compared with the traditional ID3.

Towards an automated evaluation approach for e-procurement

Automatic Procurement process is currently a vital field for research, as it is one of the basic business processes that has been applied in different fields such as construction, computer, and others. Definitely all stages in the procurement process can be more efficient when applied electronically. While different researches have focused on the automatic submission for the offers, however, in this research, a new idea for providing an evaluation process for selecting the best offer without human interference is proposed. The research provides a decision with considering many constraints with no human interaction. The current status of the provided approach is considered as supportive tool for the technical committee as decision support tool until it is further developed and evaluated.