Wyniki wyszukiwania - BazTech

1

Data mining approach in diagnosis and treatment of chronic kidney disease

Turiac Andreea S., Zdrodowska Małgorzata

Acta Mechanica et Automatica

|

2022

|

Vol. 16, no. 3

180--188

EN

Chronic kidney disease is a general definition of kidney dysfunction that lasts more than 3 months. When chronic kidney disease is advanced, the kidneys are no longer able to cleanse the blood of toxins and harmful waste products and can no longer support the proper function of other organs. The disease can begin suddenly or develop latently over a long period of time without the presence of characteristic symptoms. The most common causes are other chronic diseases – diabetes and hypertension. Therefore, it is very important to diagnose the disease in early stages and opt for a suitable treatment - medication, diet and exercises to reduce its side effects. The purpose of this paper is to analyse and select those patient characteristics that may influence the prevalence of chronic kidney disease, as well as to extract classification rules and action rules that can be useful to medical professionals to efficiently and accurately diagnose patients with kidney chronic disease. The first step of the study was feature selection and evaluation of its effect on classification results. The study was repeated for four models – containing all available patient data, containing features identified by doctors as major factors in chronic kidney disease, and models containing features selected using Correlation Based Feature Selection and Chi-Square Test. Sequential Minimal Optimization and Multilayer Perceptron had the best performance for all four cases, with an average accuracy of 98.31% for SMO and 98.06% for Multilayer Perceptron, results that were confirmed by taking into consideration the F1-Score, for both algorithms was above 0.98. For all these models the classification rules are extracted. The final step was action rule extraction. The paper shows that appropriate data analysis allows for building models that can support doctors in diagnosing a disease and support their deci-sions on treatment. Action rules can be important guidelines for the doctors. They can reassure the doctor in his diagnosis or indicate new, previously unseen ways to cure the patient.

2

Selected morphotic parameters differentiating ulcerative colitis from crohn’s disease

Kasperczuk Anna

Acta Mechanica et Automatica

|

2021

|

Vol. 15, no. 4

249--253

EN

This paper presents a method that binds statistical and data mining techniques, which aims to support the decision-making process in selected diseases of the digestive system. Currently, there is no precise diagnosis for ulcerative colitis (UC) and Crohn's disease (CD). Specialist physicians must exclude many other diseases occurring in the colon. The first goal of this study is a retrospective analysis of medical data of patients hospitalised in the Department of Gastroenterology and Internal Diseases, Bialystok, and finding the symptoms differentiating the two analysed diseases. The second goal is to build a system that clearly points to one of the two diseases UC or CD, which shortens the time of diagnosis and facilitates the future treatment of patients. The work focuses on building a model that can be the basis for the construction of action rules, which are one of the basic elements in the medical recommendation system. Generated action rules indicated differentiating factors, such as mean corpuscular volume, platelets (PLTs), neutrophils, monocytes, eosinophils, basophils, alanine aminotransferase (ALAT), creatinine, sodium and potassium. Other important parameters were smoking and blood in stool.

3

Action Rules of Lowest Cost and Action Set Correlations

Tzacheva Angelina A., Shankar Ramya A., Ramachandran Sridharan, Bagavathi Arunkumar

Fundamenta Informaticae

|

2020

|

Vol. 172, nr 4

399--412

EN

A knowledge discovery system is prone to yielding plenty of patterns, presented in the form of rules. Sifting through to identify useful and interesting patterns is a tedious and time consuming process. An important measure of interestingness is: whether or not the pattern can be used in the decision making process of a business to increase profit. Hence, actionable patterns, such as action rules, are desirable. Action rules may suggest actions to be taken based on the discovered knowledge. In this way contributing to business strategies and scientific research. The large amounts of knowledge in the form of rules presents a challenge of identifying the essence, the most important part, of high usability. We focus on decreasing the space of action rules through generalization. In this work, we present a new method for computing the lowest cost of action rules and their generalizations. We discover action rules of lowest cost by taking into account the correlations between individual atomic action sets.

4

Data mining techniques as a tool in neurological disorders diagnosis

Zdrodowska M., Dardzińska A., Chorąży M., Kułakowska A.

Acta Mechanica et Automatica

|

2018

|

Vol. 12, no. 3

217--220

EN

Neurological disorders are diseases of the brain, spine and the nerves that connect them. There are more than 600 diseases of the nervous system, such as epilepsy, Parkinson's disease, brain tumors, and stroke as well as less familiar ones such as multiple sclerosis or frontotemporal dementia. The increasing capabilities of neurotechnologies are generating massive volumes of complex data at a rapid pace. Evaluating and diagnosing disorders of the nervous system is a complicated and complex task. Many of the same or similar symptoms happen in different combinations among the different disorders. This paper provides a survey of developed selected data mining methods in the area of neurological diseases diagnosis. This review will help experts to gain an understanding of how data mining techniques can assist them in neurological diseases diagnosis and patients treatment.

5

Algorithm for generalization of action rules to summaries

Tzacheva A. A.

Control and Cybernetics

|

2010

|

Vol. 39, no 2

457-468

EN

A knowledge discovery system is prone to yielding plenty of patterns, presented in the form of rules. Sifting through to identify useful and interesting patterns is a tedious and time consuming process. An important measure of interestingness is: whether or not the pattern can be used in the decision making process of a business to increase profit. Hence, actionable patterns, such as action rules, are desirable. Action rules may suggest actions to be taken based on the discovered knowledge. In this way contributing to business strategies and scientific research. The large amounts of knowledge in the form of rules presents a challenge of identifying the essence, the most important part, of high usability. We focus on decreasing the space of action rules through generalization. In this paper, we propose an improved method for discovering short descriptions of action rules. The new algorithm produces summaries by maximizing the diversity of rule pairs, and minimizing the cost of the suggested actions.

6

Tree-based Construction of Low-cost Action Rules

Tzacheva A.A., Tsay L-S.

Fundamenta Informaticae

|

2008

|

Vol. 86, nr 1-2

213-225

EN

A rule is actionable, if a user can do an action to his/her advantage based on that rule. Actionability can be expressed in terms of attributes that are present in a database. Action rules are constructed from certain pairs of classification rules, previously extracted from the same database, each defining a preferable decision class. It is assumed that attributes are divided into two groups: stable and flexible. Flexible attributes provide a tool for making hints to a user to what changes within some values of flexible attributes are needed to re-classify a group of objects, supporting the action rule, from one decision class to another, more desirable, one. Changes of values of some flexible attributes can be more expensive than changes of other values. To investigate such cases, the notion of a cost is introduced and it is assigned by an expert to each such a change. Action rules construction involves both flexible and stable attributes listed in certain pairs of classification rules. The values of stable attributes are used to create action forest. We propose a new strategy which combines the action forest algorithm of extracting action rules and a heuristic strategy for generating reclassification rules of the lowest cost. This new strategy presents an enhancement to both methods.