پایه و اساس سیستم هوشمند ترکیبی محاسباتی و نظری مجموعه های ناهموار برای تجزیه و تحلیل بقا
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|5556||2008||10 صفحه PDF||سفارش دهید||5436 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Computers & Mathematics with Applications, Volume 56, Issue 7, October 2008, Pages 1699–1708
What do we (not) know about the association between diabetes and survival time? Our study offers an alternative mathematical framework based on rough sets to analyze medical data and provide epidemiology survival analysis with risk factor diabetes. We experiment on three data sets: geriatric, melanoma and Primary Biliary Cirrhosis. A case study reports from 8547 geriatric Canadian patients at the Dalhousie Medical School. Notification status (dead or alive) is treated as the censor attribute and the time lived is treated as the survival time. The analysis result illustrates diabetes is a very significant risk factor to survival time in our geriatric patients data. This paper offers both theoretical and practical guidelines in the construction of a rough sets hybrid intelligent system, for the analysis of real world data. Furthermore, we discuss the potential of rough sets, artificial neural networks (ANNs) and frailty index in predicting survival tendency.
Survival analysis  is a branch of statistics that studies time-to-event data. Death or failure is called an event in the survival analysis literature. Survival analysis attempts to answer questions such as: is diabetes a significant risk factor for geriatric patients? What is the fraction of patients who will survive past a certain time? Survival analysis is called reliability analysis in engineering, and duration analysis in economics. Presently, survival data in existence worldwide highlights the need for further comprehensive and systematic analysis to improve overall health outcomes. Much data analysis research has been conducted in several areas , ,  and . The aim of such data analysis techniques is to use the collected data for training in a learning process, and then to extract a hidden pattern by model construction. However, a successful technique involves far more than selecting a learning algorithm and running it over data sets. Successful data analysis requires in-depth knowledge of data. The challenges in real world problems are the complexity and unique properties of the survival data at hand. In many practical situations, survival data sets are vague and come with redundant and irrelevant attributes. The inclusion of these attributes in the data causes some difficulties in discovering the knowledge. To avoid these troubles, it is essential to precede the learning task with an attribute selection process to delete redundancy records, uncertainty attributes and overwhelming data. To this end, we create an attribute subset large enough to include all of the important attributes, but small enough for our learning system to handle easily. Another issue in survival data analysis is the desire for automatic analysis processes . Classical approaches are designed theoretically, automation is then increasingly challenging. Traditional data analysis is not adequate (e.g., Dempster–Shafer theory, grade of membership ), and methods for efficient mathematical and computer-based analysis, e.g., rough sets, are indispensable. Rough set theory was developed by Zdzislaw Pawlak , , ,  and . It provides system designers with the ability to compute with imperfect data. If a concept cannot be defined in a given knowledge base (vagueness), rough sets can approximate that knowledge efficiently. While logic is deductive and hardly applies to real situations, rough sets is in the form of inductive reasoning that widens the scope of the research to deal with real world data . Rough sets do not require a specific model that can fit the data to be used in the analysis process. This ability provides flexibility in real situations. Rough sets provide a semi-automatic approach to data analysis and can be combined with other complementary techniques. Thus, current research tends to hybridize diverse methods of soft computing . In this paper, we offer an approach based on rough sets with the capability to reason and to distil useful knowledge for survival data (e.g., risk factor, survival prediction model). This article is organized as follows. We introduce in Section 2 preliminaries of rough sets, relational algebra and other scientific areas along with hybridization of these approaches. In Section 3, we propose our rough sets hybrid intelligent system and new CDispro algorithm. We demonstrate the applicability of our system by several experiments over a range of data sets reported in Section 4. In Section 5, the evaluation results are presented and also a brief comparison to another case studies in Sections 6 and 7. We conclude in Section 8.
نتیجه گیری انگلیسی
Starting from mathematical rough set theory the central theme of this study is to invent the hybrid intelligent system. Our rough sets hybrid intelligent system is useful for survival analysis and extracting the most informative and useful knowledge. We created our system to have the following features. Our system was designed to provide comprehensive survival data analysis tasks; preprocessing, analyzing process and postprocessing. We amalgamated rough sets and other techniques in soft computing to be able to make the analyzing process tolerant to imprecise and uncertain data. We ensured the correctness of rules by designing automatic validation processes. Furthermore, the computation times were improved significantly by using database operations. The experimental results show how our rough sets hybrid intelligent system could be employed to quickly process. Clinical diagnosis questions can be answered successfully, e.g., is diabetes a significant factor for survival time of geriatric patients? Analysis results show that it has significant impact on the survival time of geriatric patients. Decision rules described particular tendencies for survival outcomes of patients by using decision rules that are straightforward and simple to use. In the future, from theoretical viewpoint, we will pay more attention to many advances in rough sets e.g., rough mereology, rough inclusion, decision logic or dissimilarly analysis. Our results offer alternative choices to the patients or anyone concerned by the outcome of medical treatments or the progression of diseases. We have illustrated that pursuing further research in this relatively young area of mathematics, rough set theory, is a worthwhile aim.