Search In this Thesis
   Search In this Thesis  
العنوان
Drug Design and Discovery Framework Based on Soft Computing Techniques /
المؤلف
Mohamed, Mosa Elkhedr Hosney.
هيئة الاعداد
باحث / موسى الخضر حسنى محمد
مشرف / عصام حليم حسين
مشرف / محمود حسب الله محمود
مشرف / وليد مكرم محمد
الموضوع
Pharmaceutical chemistry - Data processing. Drugs - Design - Data processing.
تاريخ النشر
2020.
عدد الصفحات
124 p. :
اللغة
الإنجليزية
الدرجة
ماجستير
التخصص
Information Systems
تاريخ الإجازة
1/1/2020
مكان الإجازة
جامعة المنيا - كلية الحاسبات والمعلومات - نظم المعلومات
الفهرس
Only 14 pages are availabe for public view

from 150

from 150

Abstract

The revolution of the information system is led to more integrated with other science especially chemical and medical. Cheminformatics is the use of computer and information techniques that are applied for many problems in the chemistry field. Cheminformatics has many branches but drug design and discovery are the most important of them.It is represented as a simple rational design or rational drug design. Some processes should be followed to find new medications based on biological target knowledge.The drug is an organic small molecule that inhibits or activates biomolecule function as a protein.Drugs can be indicated through computer-aided drug design discovery tools.Several parameters should be considered during designing the drug as drugs affect life so the best drug should be proposed.
Swarm algorithms and machine learning algorithms are important trends in computer science as they can simulate the behaviour of natural so these can be used for predicting and classifying various problems for improving the several chemical compound activities to propose the best drug.Several methods are introduced for improving the metaheuristics algorithms by integration with other algorithms to reflect the best result for the introduced chemical data. The proposed algorithm is improved by hybrid Harris hawks optimization with SVM and also Harris hawks optimization-genetics algorithm with k-NN to overcome the major problems of HHO that are trapped in local optimal and achieved more balancing between exploration and exploitation.Firstly, features are reduced and the best features are extracted by using wrapper feature selection.Feature selection is a preprocessing step for the classification machine learning algorithm. SVM and k-NN are used for the proposed algorithm.The best accuracy results are showed by our approach for chemical data set compared with several well-established metaheuristic algorithms as Particle Swarm Optimization (PSO), Simulated Annealing (SA), Dragonfly Algorithm (DA), Butterfly Optimization Algorithm (BOA), Moth-Flame Optimization Algorithm (MFO), Grey Wolf Optimizer (GWO), Sine Cosine Algorithm (SCA),Slap Swarm Algorithm (SSA),Ant Lion Optimizer (ALO) and Whale Optimization Algorithm (WOA) so in first case,the optimal HHO-SVM result is 97.58 % for MOA and 85.023% to QSAR Biodegradation and also in the second case, HHO-CM with k-NN is 99.52% for MOA data set and 90.30 % to QSAR Biodegradation.