Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Decision Trees”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 37 records · Page 2Linked to original sources
PubMed2005

Classifiability-based omnivariate decision trees.

Top-down induction of decision trees is a simple and powerful method of pattern classification. In a decision tree, each node partitions the available patterns into two or more sets. New nodes are created to handle each of the resulting partitions and the process continues. A node is considered terminal if it satisfies some stopping criteria (for example, purity, i.e., all patterns at the node are from a single class). Decision trees may be univariate, linear multivariate, or nonlinear multivariate depending on whether a single attribute, a linear function of all the attributes, or a nonlinear function of all the attributes is used for the partitioning at each node of the decision tree. Though nonlinear multivariate decision trees are the most powerful, they are more susceptible to the risks of overfitting. In this paper, we propose to perform model selection at each decision node to build omnivariate decision trees. The model selection is done using a novel classifiability measure that captures the possible sources of misclassification with relative ease and is able to accurately reflect the complexity of the subproblem at each node. The proposed approach is fast and does not suffer from as high a computational burden as that incurred by typical model selection algorithms. Empirical results over 26 data sets indicate that our approach is faster and achieves better classification accuracy compared to statistical model select algorithms.

Algorithms↗
PubMed2002

Predicting recovery in patients suffering from traumatic brain injury by using admission variables and physiological data: a comparison between decision tree analysis and logistic regression.

OBJECT: Decision tree analysis highlights patient subgroups and critical values in variables assessed. Importantly, the results are visually informative and often present clear clinical interpretation about risk factors faced by patients in these subgroups. The aim of this prospective study was to compare results of logistic regression with those of decision tree analysis of an observational, head-injury data set, including a wide range of secondary insults and 12-month outcomes. METHODS: One hundred twenty-four adult head-injured patients were studied during their stay in an intensive care unit by using a computerized data collection system. Verified values falling outside threshold limits were analyzed according to insult grade and duration with the aid of logistic regression. A decision tree was automatically produced from root node to target classes (Glasgow Outcome Scale [GOS] score). Among 69 patients, in whom eight insult categories could be assessed, outcome at 12 months was analyzed using logistic regression to determine the relative influence of patient age, admission Glasgow Coma Scale score, Injury Severity Score (ISS), pupillary response on admission, and insult duration. The most significant predictors of mortality in this patient set were duration of hypotensive, pyrexic, and hypoxemic insults. When good and poor outcomes were compared, hypotensive insults and pupillary response on admission were significant. Using decision tree analysis, the authors found that hypotension and low cerebral perfusion pressure (CPP) are the best predictors of death, with a 9.2% improvement in predictive accuracy (PA) over that obtained by simply predicting the largest outcome category as the outcome for each patient. Hypotension was a significant predictor of poor outcome (GOS Score 1-3). Low CPP, patient age, hypocarbia, and pupillary response were also good predictors of outcome (good/poor), with a 5.1% improvement in PA. In certain subgroups of patients pyrexia was a predictor of good outcome. CONCLUSIONS: Decision tree analysis confirmed some of the results of logistic regression and challenged others. This investigation shows that there is knowledge to be gained from analyzing observational data with the aid of decision tree analysis.

Adult↗
PubMed1980

Decision trees: construction, uses, and limits.

Decision trees are models of the temporal and logical flow of clinical problems. Their purpose is to help the physician choose a clinical management strategy that offers the greatest expected value for the patient. Decision trees help answer questions such as:"should a risky diagnostic test be performed?" "Given our present knowledge, which of several available treatments is best for this patient?" "What are the expected benefits, risks, and financial costs of pursuing different stages of patient care?" Decision trees do not create new information, but they can provide new insights based on existing information. The principles of analyzing a clinical situation from a decision analytic perspective and of constructing and using a decision tree are illustrated with three clinical examples: a patient with possible urinary tract infection, a young man with Hodgkin's disease, and patients with chronic progressive liver failure. We present a simplified quantitative analysis to determine whether the patient with Hodgkin's disease should undergo a staging laparotomy. The last example serves as a prelude to R. Fuhrer's discussion of the expected value of test information. Following R. Fuhrer's presentation, we discuss some of the objections and advantages to medical decision analysis. Despite its limitations we believe decision analysis can be a powerful aid to medical practitioners.

Adolescent↗
PubMed2001

Using decision tree induction to model oculomotor data.

Decision tree induction is a machine learning method used to generate classification models from data sets. Numerous decision trees were constructed to examine relationships between oculomotor test parameters and lesion sites in a data set containing cases with operated cerebello-pontine angle tumour, operated hemangioblastoma, infarction of cerebello-brainstem and Ménière's disease, and control subjects. The aim was to find useful parameter combinations with discriminatory power. Decision trees constructed using both pursuit eye movements and saccadic eye movements yielded the best classification results. This is reasonable: oculomotor test results vary according to the site of the lesion and so the performance ability of subjects has to be taken into account in the classification. The decision tree program was able to generate classification models from the oculomotor data set. Generated decision trees were intelligible and can be utilized in physicians' research work.

Decision Making↗
PubMedDate not supplied

Using decision trees to aid decision-making in nursing.

This article discusses judgement and decision-making in nursing. It outlines an approach to analysing clinical problems known as decision analysis. It suggests decision analysis can be a useful technique for nurses to assist them with decision-making in practice.

Decision Trees↗
PubMed2004

Developing a decision tree algorithm for the diagnosis of suspected spider bites.

OBJECTIVE: To develop a diagnostic algorithm (decision tree) to improve the ability to identify or predict medically important spider bites (funnel-web and redback spiders) from information about the circumstances and initial clinical effects of spider bites. METHODS: A dataset of definite spider bites with expert identification of all spiders was used from a previous Australia-wide prospective study. Spider bites were categorized as: big black spider (BBS), redback spider (RED) and other spider (OTH). Big black spider included funnel-web spiders (most medically significant), but also other spiders of similar appearance. Fifteen predictor variables were based on univariate analysis from previous studies and clinical experience. They included information about the circumstances and early clinical effects of bites. The data were analyzed using CART (Classification and Regression Trees), a 'decision tree' algorithm used to create a tree-like structure to describe a data set. RESULTS: Of 789 spider bites there were 49 (6.2%) bites by BBS, 68 (8.6%) bites by RED and 672 (85.2%) bites by OTH. A decision tree was developed that included six predictor variables (fang marks/bleeding; state/territory; local diaphoresis; month; time of day; and proximal or distal bite region). The decision tree accurately classified 47 out of the 49 (96%) BBS, and no funnel-web spiders were incorrectly classified (100% sensitivity). Two hundred and forty-four of 789 were classified as OTH and included no BBS. CONCLUSIONS: A decision tree based on a small amount of information about the circumstances and early clinical effects of spider bites safely predicted all funnel-web spider bites. Application of this algorithm would allow the early institution of appropriate treatment for funnel-web spider bites and the immediate discharge of 31% as other spider bites (reassurance only).

Algorithms↗
PubMed2001

Comparison of three databases with a decision tree approach in the medical field of acute appendicitis.

Decision trees have been successfully used for years in many medical decision making applications. Transparent representation of acquired knowledge and fast algorithms made decision trees one of the most often used symbolic machine learning approaches. This paper concentrates on the problem of separating acute appendicitis, which is a special problem of acute abdominal pain from other diseases that cause acute abdominal pain by use of an decision tree approach. Early and accurate diagnosing of acute appendicitis is still a difficult and challenging problem in everyday clinical routine. An important factor in the error rate is poor discrimination between acute appendicitis and other diseases that cause acute abdominal pain. This error rate is still high, despite considerable improvements in history-taking and clinical examination, computer-aided decision-support and special investigation, such as ultrasound. We investigated three different large databases with cases of acute abdominal pain to complete this task as successful as possible. The results show that the size of the database does not necessary directly influence the success of the decision tree built on it. Surprisingly we got the best results from the decision trees built on the smallest and the biggest database, where the database with medium size (relative to the other two) was not so successful. Despite that we were able to produce decision tree classifiers that were capable of producing correct decisions on test data sets with accuracy up to 84%, sensitivity to acute appendicitis up to 90%, and specificity up to 80% on the same test set.

Abdomen, Acute↗
PubMedDate not supplied

Methods of decision analysis: protocols, decision trees, and algorithms in medicine.

Algorithms, decision trees, and protocols are defined and explained since they constitute an accepted part of clinical decision analysis and application to clinical care. Algorithms are particularly useful for common clinical problems where uncertainties are unlikely. Decision trees are helpful when--as usually occurs in difficult clinical decisions--there are problems in probability. Clinical protocols, which, at best, are based on algorithms and decision trees, provide instruction of how to best treat a patient given the strict definitions of the clinical problem. These techniques are, in essence, merely graphic representations of a logical scientific approach to clinical problems. Criticisms of these techniques center on their rigidity and the automatic unthinking cookbook medicine they might sponsor. It is concluded that if these techniques are wisely designed and, even more importantly, wisely administered with an understanding flexibility, they can lead to both economy and patient benefit.

Algorithms↗
PubMed1998

Impact of a decision tree on chronic wound care.

OBJECTIVE: To examine the use of a decision tree and its impact on the accuracy of decision making for chronic wound care. DESIGN: Descriptive comparative design; replication study. SETTINGS AND SUBJECTS: Data were collected from 2 groups of home care nurses in large urban centers. One group was measured after initial contact with the decision tree, and the other group was measured 2 years after implementation of the decision tree. INSTRUMENTS: The chronic wound management decision tree (CWMDT) designed by Knight was used, in combination with pictorial case studies developed by Willey and Swords. METHODS: Edmonton Home Care Program nurses, having used a decision tree for 2 years, completed 3 pictorial case studies without the aid of a written copy of the CWMDT. Current accuracy of wound care staging and treatment was compared with the Melchior-MacDougall and Lander results. After this procedure, Calgary Health Services home care nurses completed 3 pictorial case studies to assess the accuracy of wound care staging and treatment between 2 groups: with the CWMDT and without the CWMDT. A questionnaire was completed on factors identified to influence decision making in wound care. MAIN OUTCOME MEASURE: Accuracy and confidence of decision making in wound care staging and treatment. RESULTS: Accuracy of decision making with use of the CWMDT improved over time, as did accuracy of decision making after initial contact with the decision tree. Age, experience in nursing and home care, number of in-services attended, and number of chronic wounds treated, were not found to correlate with accuracy of decision making. CONCLUSION: A decision tree can assist with decision making by guiding the nurse through assessment and treatment options.

Adult↗
PubMed2005

Boosting alternating decision trees modeling of disease trait information.

We applied the alternating decision trees (ADTrees) method to the last 3 replicates from the Aipotu, Danacca, Karangar, and NYC populations in the Problem 2 simulated Genetic Analysis Workshop dataset. Using information from the 12 binary phenotypes and sex as input and Kofendrerd Personality Disorder disease status as the outcome of ADTrees-based classifiers, we obtained a new quantitative trait based on average prediction scores, which was then used for genome-wide quantitative trait linkage (QTL) analysis. ADTrees are machine learning methods that combine boosting and decision trees algorithms to generate smaller and easier-to-interpret classification rules. In this application, we compared four modeling strategies from the combinations of two boosting iterations (log or exponential loss functions) coupled with two choices of tree generation types (a full alternating decision tree or a classic boosting decision tree). These four different strategies were applied to the founders in each population to construct four classifiers, which were then applied to each study participant. To compute average prediction score for each subject with a specific trait profile, such a process was repeated with 10 runs of 10-fold cross validation, and standardized prediction scores obtained from the 10 runs were averaged and used in subsequent expectation-maximization Haseman-Elston QTL analyses (implemented in GENEHUNTER) with the approximate 900 SNPs in Hardy-Weinberg equilibrium provided for each population. Our QTL analyses on the basis of four models (a full alternating decision tree and a classic boosting decision tree paired with either log or exponential loss function) detected evidence for linkage (Z >or= 1.96, p < 0.01) on chromosomes 1, 3, 5, and 9. Moreover, using average iteration and abundance scores for the 12 phenotypes and sex as their relevancy measurements, we found all relevant phenotypes for all four populations except phenotype b for the Karangar population, with suggested subgroup structure consistent with latent traits used in the model. In conclusion, our findings suggest that the ADTrees method may offer a more accurate representation of the disease status that allows for better detection of linkage evidence.

Algorithms↗
PubMed1989

[Construction of a decision tree for preoperative diagnosis in urinary incontinence].

In order to analyse the decisions used in non-urodynamic preoperative diagnosis of female urinary incontinence we constructed a decision tree, consisting of decision nodes and probability nodes. Each branch of the decision tree leads to an end point (terminal node), to which an utility must be assigned. We describe the different steps in the construction of the decision tree: 1. listing of all possible decisions (questionnaire, clinical stress test, both, no preoperative tests), 2. assignment of probabilities to the different events (e.g. cure, failure, operation in the absence of genuine stress incontinence), and 3. assignment of utility factors to the end points. We used previously published probabilities from our clinic, utility was assigned arbitrarily using a score from -10 to +90. The decision tress allows not only the calculation of the strategy with the highest expected utility, but also threshold analysis, sensitivity analysis and cost-benefit analysis. We feel that decision trees are valuable tools for analysing how medical decisions are reached and are particularly useful in teaching situations.

Adult↗
PubMed2005

Genetically optimized fuzzy decision trees.

In this study, we are concerned with genetically optimized fuzzy decision trees (G-DTs). Decision trees are fundamental architectures of machine learning, pattern recognition, and system modeling. Starting with the generic decision tree with discrete or interval-valued attributes, we develop its fuzzy set-based generalization. In this generalized structure we admit the values of the attributes that are represented by some membership functions. Such fuzzy decision trees are constructed in the setting of genetic optimization. The underlying genetic algorithm optimizes the parameters of the fuzzy sets associated with the individual nodes where they play a role of fuzzy "switches" by distributing a flow of processing completed within the tree. We discuss various forms of the fitness function that help capture the essence of the problem at hand (that could be either of classification nature when dealing with discrete outputs or regression-like when handling a continuous output variable). We quantify a nature of the generalization of the tree by studying an optimally adjusted spreads of the membership functions located at the nodes of the decision tree. A series of experiments exploiting synthetic and machine learning data is used to illustrate the performance of the G-DTs.

Algorithms↗
PubMed1999

Immunophenotypic diagnosis of acute leukemia by using decision tree induction.

We describe a model for decision making for bone marrow immunophenotypic analysis of acute leukemia. In this study, we used decision tree induction as an information processing system for the analysis of flow cytometry immunophenotype results of bone marrow specimens obtained for the diagnosis of acute leukemia. By using decision tree analysis, we queried which antibodies and at what percentage cut-offs led to particular diagnoses. Flow cytometry results of up to 27 monoclonal antibodies from bone marrow specimens of 175 adult and pediatric cases were used: acute lymphoblastic leukemia (n = 80), myeloid leukemia (n = 44), mixed lineage (n = 16), and reactive marrow (n = 35). The percentage of positive cells was used as input data, and the diagnoses were used as output of the information processing system. Results of the decision tree showed an easy, accurate, and intuitive algorithm that can delineate a hierarchy of antibodies relevant to diagnosis. A correct discrimination of acute myeloid and lymphoid leukemia from benign bone marrow can be inferred by using the results of four to eight from a panel of up to 27 antibodies with an accuracy of 95%. Here, we describe a computer-aided model that uses decision tree induction applied to flow cytometry immunophenotype data. If generalizable, this technique may be an alternative approach to modeling complex information like that seen in hematopathology and may complement the immunologist's interpretation, along with cytochemistry and morphology results, in the diagnosis of acute leukemia.

Acute Disease↗
PubMed1995

ROSE: decision trees, automatic learning and their applications in cardiac medicine.

Computerized information systems, especially decision support systems, have acquired an increasingly important role in medical applications, particularly in those where important decisions must be made effectively and reliably. But the possibility of using computers in medical decision making is limited by many difficulties, including the complexity of conventional computer languages, methodologies, and tools. Thus a conceptual simple decision making model with the possibility of automating learning should be used. In this paper, we introduce a cardiological knowledge-based system based on the decision tree approach supporting the mitral valve prolapse determination. Prolapse is defined as the displacement of a bodily part from its normal position. The term mitral valve prolapse (PMV), therefore, implies that the mitral leaflets are displaced relative to some structure, generally taken to be the mitral annulus. The implications of the PMV are: disturbed normal laminar blood flow, turbulence of the blood flow, injury of the chordae tendinae, the possibility of thrombus's composition, bacterial endocarditis, and, finally, hemodynamic changes defined as mitral insufficiency and mitral regurgitation. Uncertainty persists about how it should be diagnosed and about its clinical importance. It is our deep belief that the echocardiography enables properly trained expert armed with proper criteria to evaluate PMV almost 100%. But, unfortunately, there are some problems concerned with the use of echocardiography. With this in mind, we have decided to start a research project aimed at finding new criteria and enabling the general practitioner to evaluate the PMV using conventional methods and to select potential patients from the general population. To empower doctors to perform needed activities, we have developed a computer tool called ROSE (computeRized prOlaps Syndrome dEtermination) based on algorithms of automatic learning. This tool supports the definition of new criteria and the selection of potential PMV-patients. The ROSE is based on concepts of decision trees and automatic learning. The decisions and learning process can be presented with an easily visualized two dimensional model; thus decision trees are straightforward to build and interpret. Decision trees use different object attributes to classify different subsets of objects. (Their great advantage is that they don't use a fixed number of predetermined attributes.) In the decision tree approach, the members of a set of objects are classified as either positive or negative instances (in our case patients with PMV Syndrome or without it). Candidate attributes that may possibly describe the concept are then outlined. A decision tree construction tool uses outlined attributes to formulate the appropriate decision tree that identifies all positive instances of the underlying concept according to objects with known classification. (In our case the classification is done with the echo examination). The first set of objects used for the tree generation is usually called the training set. This decision tree characterization next becomes a basis: 1) forecasting whether an object previously unseen is a positive or negative instance of the concept being modeled, and 2) the hierarchical representation of the most important attributes of the concept being investigated. Our main interest and idea is to discover symptoms, syndromes, and illnesses related to PMV that can be distinguished by general practitioners in their everyday job and which should help them to identify possible PMV candidate patients. To this end, we constructed a computerized tool called ROSE. According to the principles presented above, we first taught ROSE using the sample of 400 examined volunteers. ROSE, considering that the clinical PMV diagnosis is practically not researched, is relatively successful. (abstract truncated)

Algorithms↗
PubMed2004

Measuring performance in health care: case-mix adjustment by boosted decision trees.

OBJECTIVE: The purpose of this paper is to investigate the suitability of boosted decision trees for the case-mix adjustment involved in comparing the performance of various health care entities. METHODS: First, we present logistic regression, decision trees, and boosted decision trees in a unified framework. Second, we study in detail their application for two common performance indicators, the mortality rate in intensive care and the rate of potentially avoidable hospital readmissions. RESULTS: For both examples the technique of boosting decision trees outperformed standard prognostic models, in particular linear logistic regression models, with regard to predictive power. On the other hand, boosting decision trees was computationally demanding and the resulting models were rather complex and needed additional tools for interpretation. CONCLUSION: Boosting decision trees represents a powerful tool for case-mix adjustment in health care performance measurement. Depending on the specific priorities set in each context, the gain in predictive power might compensate for the inconvenience in the use of boosted decision trees.

Decision Trees↗
PubMedDate not supplied

[Decision trees in psychiatric therapy].

The main objective of decision analysis is to offer a theoretical representation of choices made in an environment of uncertainty. This technique is currently under development in a great variety of fields, particularly in medicine, where aid in decision making is the topic of much research. Psychiatry, in turn, is very much concerned by these new developments which could be of particular interest to therapeutics-an area where the large number of studies and date are in great contrast with the lack of consensus concerning the various solutions proposed to patients. Decision analysis utilizes different techniques among which are decision trees. The technique of decision trees goes far beyond a simple graphic representation of reasoning in the form of a chart. Its basic principles is to measure the uncertainty associated with decision making in the hopes of better understanding the rationale of decisions while optimizing the gain versus cost ratio. The goal is to calculate, within a series of decisions, the weight of their importance expressed in terms of usefulness or unpleasantness. In psychiatric therapeutics, only three studies have been published which incorporate the technique of decision trees. Two of these deal with treating depression (Schulberg et al., 1989; Koenig et al., 1993) while the third deals with schizophrenia (Hatcher, 1995). The limits of these techniques are, on one hand, due to their feasibility in that their complexity renders them inapplicable when a great number of variables have to be taken into account or when the amount of necessary data is still insufficient. Moreover, the use of these techniques remains relatively restricted as their expansion depends upon their acceptance by clinical physicians. Also, their use raises questions as to what extent it is possible to rationalize decisions in psychiatry. From a larger perspective, one must consider that these techniques may eventually furnish certain elements which could be integrated to help further the field of decision-making representations for clinical use. These decision-making techniques are still in the experimental stages and remains difficult to apply to clinical practice. However they appear to be of a great interest, not only in communicating knowledge both in teaching and training, but in research as well. They allow us to view the results of epidemiological studies and clinical research from a more global perspective; to make evident the grey areas of our science and to determine new priorities in research.

Decision Trees↗
PubMed2006

Decision tree methods in pharmaceutical research.

Decision trees are among the most popular of the new statistical learning methods being used in the pharmaceutical industry for predicting quantitative structure-activity relationships. This article reviews applications of decision trees in drug discovery research and extensions to the basic algorithm using hybrid or ensemble methods that improve prediction accuracy.

Chemistry, Pharmaceutical↗
PubMed2001

Finding the right decision tree's induction strategy for a hard real world problem.

Decision trees have been already successfully used in medicine, but as in traditional statistics, some hard real world problems can not be solved successfully using the traditional way of induction. In our experiments we tested various methods for building univariate decision trees in order to find the best induction strategy. On a hard real world problem of the Orthopaedic fracture data with 2637 cases, described by 23 attributes and a decision with three possible values, we built decision trees with four classical approaches, one hybrid approach where we combined neural networks and decision trees, and with an evolutionary approach. The results show that all approaches had problems with either accuracy, sensitivity, or decision tree size. The comparison shows that the best compromise in hard real world problem decision trees building is the evolutionary approach.

Algorithms↗