Recognition of the respiratory pathology by statistic methods

B. Sokolowska, K. Budzinska, A. Jozwik (Warsaw, Poland)

Source: Annual Congress 2001 - Monitoring acute lung disease
Session: Monitoring acute lung disease
Session type: Thematic Poster Session
Number: 1650
Disease area: Respiratory critical care

Congress or journal article abstract

Abstract

The experiments were performed on 12 anesthetized, tracheostomized, and spontaneously breathing adult cats before and 1 hour after bilateral phrenicotomy. Three parameters: frequency of breathing f, tidal volume VT and minute ventilation VE were measured during resting ventilation. Measurements of features were repeated 9 times for each the animal in the two states of the respiratory system. The data set contained two subsets by 108 of 3-dimensional measurement vectors. The question was if it possible to recognize on the basis of the collected data whether the cats breathed with or without the diaphragm. To study this issue the methods of statistical pattern recognition theory were applied. With these methods decision rules were constructed that allowed evaluating the considered parameters in their mutual context. Four different types of the decision rules were used and compared: the standard k-nearest neighbor rule (k-NN), fuzzy version of k-NN rule, standard k-NN rule with a preclassifier and the fuzzy k-NN rule with the preclassifier. For each of these decision rules all parameter combinations were analyzed. The highest probability of correct classification, equal 98%, was offered by the fuzzy k-NN rule with the preclassifier and the 2 parameters f and VT. This high probability indicates the strong influence of diaphragm denervation on the ventilation parameters f and VT. The parameter VE was redundant and its use decreased the probability of correct classification by 1%, i.e. to the value of 97%. The results show that presented approach based on statistical pattern recognition methods is an effective tool for the medical decision making.


Rating: 0
You must login to grade this presentation.

Share or cite this content

Citations should be made in the following way:
B. Sokolowska, K. Budzinska, A. Jozwik (Warsaw, Poland). Recognition of the respiratory pathology by statistic methods. Eur Respir J 2001; 16: Suppl. 31, 1650

You must login to share this Presentation/Article on Twitter, Facebook, LinkedIn or by email.

Member's Comments

No comment yet.
You must Login to comment this presentation.


Related content which might interest you:
Boyd OODA loop and Wilcoxon rank sum test in diagnostic procedures after major disasters with lung involvement
Source: International Congress 2015 – Patient health gains related to the economic effectiveness of disease management programmes
Year: 2015

Use of diagnostic tests in respiratory tract infections: an economic analysis.
Source: Virtual Congress 2021 – Emerging data for the management of respiratory tract infections
Year: 2021


Mortality and classification of airway obstruction severity by different methods
Source: International Congress 2016 – Quality and quantity in lung function
Year: 2016


The airspace dimension test – A new way to diagnose emphysema?
Source: International Congress 2014 – Lung function testing: new findings and new approaches
Year: 2014

A new statistical diagnostic tool for respiratory diseases
Source: Virtual Congress 2020 – New and integrated tools for respiratory diagnosis and care
Year: 2020




Contribution of serological tests to the diagnosis of lower respiratory tract infections(LRTI)
Source: Eur Respir J 2004; 24: Suppl. 48, 373s
Year: 2004

Artificial intelligence outperforms pulmonologists in the interpretation of pulmonary function tests
Source: Eur Respir J, 53 (4) 1801660; 10.1183/13993003.01660-2018
Year: 2019



Explaining predictions of an automated pulmonary function test interpretation algorithm
Source: International Congress 2019 – M-health/e-health I
Year: 2019

Comparison of two different pulmonary function test (PFT) interpretive strategies?
Source: International Congress 2014 – Lung function testing: new findings and new approaches
Year: 2014

Fine crackles quantitative value can help the diagnosis of interstitial lung diseases – Clinical utility of the innovative analyzing system of respiratory sounds
Source: International Congress 2019 – Innovations in primary care assessment and management
Year: 2019

Evaluation of a method for respiratory muscle function testing
Source: Annual Congress 2007 - Lung function and therapy in obstructive lung diseases
Year: 2007


Speed up PCD diagnosis: a quick screening test for recurrent respiratory infections in children
Source: International Congress 2017 – Paediatric bronchology
Year: 2017

Comparison of pneumonia severity scoring methods in identification of severe community acquired pneumonia
Source: International Congress 2017 – CAP: severity scores, risk groups and biomarkers
Year: 2017

Airspace dimension test (ADT): A novel technique for diagnosis of the peripheral lung
Source: International Congress 2014 – Lung function testing: new findings and new approaches
Year: 2014


Repercussions to respiratory function tests
Source: Annual Congress 2009 - Difficult-to-treat tracheal pathology: pulmonological and surgical problem
Year: 2009


Comparison of scoring systems use for estimation of severity of community acquired pneumonia
Source: Annual Congress 2012 - Management of pneumonia due to hospital pathogens
Year: 2012

The evolution of respiratory function testing in paediatric lung diseases
Source: Virtual Congress 2020 – Lung health and respiratory function at all ages
Year: 2020


E-nose discriminative ability within a highly homogeneous sample of respiratory diseases
Source: International Congress 2016 – Assessing lower airway inflammation with different tools
Year: 2016


Hierarchy of common lung function parameters in diagnosing COPD using discriminant analysis
Source: Annual Congress 2010 - New diagnostic approaches in respiratory function
Year: 2010