Modelling the probability of Australian grassfires escaping initial attack to aid deployment decisions

2013 ◽  
Vol 22 (4) ◽  
pp. 459 ◽  
Author(s):  
Matt P. Plucinski
Keyword(s):  
Logistic Regression ◽  
Response Times ◽  
Classification Trees ◽  
Predictor Variables ◽  
Initial Attack ◽  
Small Areas ◽  
Long Duration ◽  
Fire Area ◽  
Weather Parameters ◽  
Operational Data

Most grassfires that occur in southern Australia are contained to small areas by local suppression resources. Those that are not require extra resources from neighbouring districts. Identifying these fires at the start of initial attack can prompt early resource requests so that resources arrive earlier when they can more effectively assist with containment. This study uses operational data collected from Australian grassfires that used ground tankers and aircraft for suppression. Variables were limited to those available when the first situation report is provided to incident controllers and included weather parameters, resource response times, slope, curing state, pasture condition and estimated fire area at initial attack. Logistic regression and classification trees were used to identify grassfires likely to escape initial attack by (a) becoming large (final area ≥100 ha), (b) being of long duration (containment time ≥4 h) or (c) either or both of these. These fires would benefit from having more resources deployed to them than are normally available. The best models used initial fire area and Grassland Fire Danger Index as predictor variables. Preliminary operational decision guides developed from classification trees could be used by fire managers to make quick assessments of the need for extra resources at early stages of a fire.

Characteristics of Adenocarcinoma of the Gallbladder Following Cholecystectomies in the Hispanic Population

2019 ◽  
Vol 152 (Supplement_1) ◽  
pp. S64-S65
Author(s):  
David Gustafson ◽  
Osvaldo Padilla
Keyword(s):  
Risk Factors ◽  
Logistic Regression ◽  
Regression Analysis ◽  
Logistic Regression Analysis ◽  
Incidental Finding ◽  
Conditional Logistic Regression ◽  
Gallbladder Disease ◽  
Predictor Variables ◽  
Wall Thickening ◽  
Conditional Logistic Regression Analysis

Abstract Introduction Gallbladder adenocarcinoma (GBC) is a rare malignancy. Frequency of incidental adenocarcinoma of the gallbladder in the literature is approximately 0.2% to 3%. Typically, GBC is the most common type and is discovered late, not until significant symptoms develop. Common symptoms include right upper quadrant pain, nausea, anorexia, and jaundice. A number of risk factors in the literature are noted for GBC. These risk factors are also more prevalent in Hispanic populations. This study sought to compare patients with incidental gallbladder adenocarcinomas (IGBC) to those with high preoperative suspicion for GBC. Predictor variables included age, sex, ethnicity, radiologic wall thickening, gross pathology characteristics (wall thickness, stone size, stone number, and tumor size), histologic grade, and staging. Methods Cases of GBC were retrospectively analyzed from 2009 through 2017, yielding 21 cases. Data were collected via Cerner EMR of predictor variables noted above. Statistical analysis utilized conditional logistic regression analysis. Results The majority of patients were female (n = 20) and Hispanic (n = 19). There were 14 IGBCs and 7 nonincidental GBCs. In contrast with previous research, exact conditional logistic regression analysis revealed no statistically significant findings. For every one-unit increase in AJCC TNM staging, there was a nonsignificant 73% reduction in odds (OR = 0.27) of an incidental finding of gallbladder carcinoma. Conclusion This study is important in that it attempts to expand existing literature regarding a rare type of cancer in a unique population, one particularly affected by gallbladder disease. Further studies are needed to increase predictive knowledge of this cancer. Longer studies are needed to examine how predictive power affects patient outcomes. This study reinforces the need for routine pathologic examination of cholecystectomy specimens for cholelithiasis.

Abstract 591: Can Waveform Characteristics Predict Defibrillation to a Perfusing Rhythm in Short Versus Prolonged Ventricular Fibrillation in a Swine Model?

Circulation ◽  
2008 ◽  
Vol 118 (suppl_18) ◽  
Author(s):  
Julia H Indik ◽  
Mathias Zuercher ◽  
Karl B Kern ◽  
Ronald W Hilwig ◽  
Robert A Berg
Keyword(s):  
Logistic Regression ◽  
Ventricular Fibrillation ◽  
Short Duration ◽  
Important Determinant ◽  
Swine Model ◽  
Median Frequency ◽  
Chest Compressions ◽  
Long Duration ◽  
Spectral Area ◽  
Minute Duration

It is known that defibrillation of ventricular fibrillation (VF) to a perfusing rhythm (ROSC) is more likely to occur in VF of short duration. It is unknown whether ROSC can be predicted by waveform characteristics in VF of short compared to long duration, apart from a consideration of time alone. VF was untreated for 2 minutes (N=10) or 8 minutes (N=10) in normal swine, after which a defibrillation shock was applied. Chest compressions for two minutes were allowed following but not prior to the shock to achieve a perfusing rhythm (ROSC). VF was analyzed from needle electrodes prior to the shock for amplitude spectral area (AMSA), slope, median frequency and bandwidth. Predictors of ROSC were determined by logistic regression. In VF of 2 minute duration 7 out of 10 swine achieved ROSC compared to 2 out 10 swine with VF of 8 minutes (P=0.025) and time was a significant predictor of ROSC (P=0.033). AMSA was significantly higher at 2 minutes (75 ± 18 mV-Hz) compared to 8 minutes (56±11 mV-Hz, p=0.007) as was slope (3.5±1 vs 2.6±0.5 mV/s, p=0.015). Bandwidth was slightly increased from 2.2±0.6 Hz at 2 minutes to 2.8±0.8 Hz at 8 minutes,(p=0.048), while median frequency was similar. However, no waveform characteristic was a significant predictor of ROSC, with substantial overlap in distributions between animals with and without ROSC. Duration of VF is an important determinant of the likelihood of achieving ROSC with defibrillation. Particularly in VF of short duration, VF waveform characteristics do not add to the predictability of achieving ROSC even though they may demonstrate a significant time evolution.

scholarly journals Predicting Hospitalization following Psychiatric Crisis Care using Machine Learning

2019 ◽  
Author(s):  
Matthijs Blankers ◽  
Louk F. M. van der Post ◽  
Jack J. M. Dekker
Keyword(s):  
Machine Learning ◽  
Logistic Regression ◽  
Learning Algorithms ◽  
Nearest Neighbors ◽  
Machine Learning Algorithms ◽  
Predictor Variables ◽  
Gradient Boosting ◽  
K Nearest Neighbors ◽  
Psychiatric Crisis ◽  
Crisis Care

Abstract Background: It is difficult to accurately predict whether a patient on the verge of a potential psychiatric crisis will need to be hospitalized. Machine learning may be helpful to improve the accuracy of psychiatric hospitalization prediction models. In this paper we evaluate and compare the accuracy of ten machine learning algorithms including the commonly used generalized linear model (GLM/logistic regression) to predict psychiatric hospitalization in the first 12 months after a psychiatric crisis care contact, and explore the most important predictor variables of hospitalization. Methods: Data from 2,084 patients with at least one reported psychiatric crisis care contact included in the longitudinal Amsterdam Study of Acute Psychiatry were used. The accuracy and area under the receiver operating characteristic curve (AUC) of the machine learning algorithms were compared. We also estimated the relative importance of each predictor variable. The best and least performing algorithms were compared with GLM/logistic regression using net reclassification improvement analysis. Target variable for the prediction models was whether or not the patient was hospitalized in the 12 months following inclusion in the study. The 39 predictor variables were related to patients’ socio-demographics, clinical characteristics and previous mental health care contacts. Results: We found Gradient Boosting to perform the best (AUC=0.774) and K-Nearest Neighbors performing the least (AUC=0.702). The performance of GLM/logistic regression (AUC=0.76) was above average among the tested algorithms. Gradient Boosting outperformed GLM/logistic regression and K-Nearest Neighbors, and GLM outperformed K-Nearest Neighbors in a Net Reclassification Improvement analysis, although the differences between Gradient Boosting and GLM/logistic regression were small. Nine of the top-10 most important predictor variables were related to previous mental health care use. Conclusions: Gradient Boosting led to the highest predictive accuracy and AUC while GLM/logistic regression performed average among the tested algorithms. Although statistically significant, the magnitude of the differences between the machine learning algorithms was modest. Future studies may consider to combine multiple algorithms in an ensemble model for optimal performance and to mitigate the risk of choosing suboptimal performing algorithms.

scholarly journals Bankruptcy Prediction of Engineering Companies in the EU Using Classification Methods

2018 ◽  
Vol 66 (5) ◽  
pp. 1347-1356 ◽  
Author(s):  
Michaela Staňková ◽  
David Hampel
Keyword(s):  
Logistic Regression ◽  
Support Vector Machines ◽  
Binary Classification ◽  
Classification Tree ◽  
Classification Trees ◽  
Bankruptcy Prediction ◽  
Support Vector ◽  
Type I ◽  
Vector Machines ◽  
The Eu

This article focuses on the problem of binary classification of 902 small- and medium‑sized engineering companies active in the EU, together with additional 51 companies which went bankrupt in 2014. For classification purposes, the basic statistical method of logistic regression has been selected, together with a representative of machine learning (support vector machines and classification trees method) to construct models for bankruptcy prediction. Different settings have been tested for each method. Furthermore, the models were estimated based on complete data and also using identified artificial factors. To evaluate the quality of prediction we observe not only the total accuracy with the type I and II errors but also the area under ROC curve criterion. The results clearly show that increasing distance to bankruptcy decreases the predictive ability of all models. The classification tree method leads us to rather simple models. The best classification results were achieved through logistic regression based on artificial factors. Moreover, this procedure provides good and stable results regardless of other settings. Artificial factors also seem to be a suitable variable for support vector machines models, but classification trees achieved better results using original data.

On quadratic logistic regression models when predictor variables are subject to measurement error

2016 ◽  
Vol 95 ◽  
pp. 109-121 ◽  
Author(s):  
Jakub Stoklosa ◽  
Yih-Huei Huang ◽  
Elise Furlan ◽  
Wen-Han Hwang
Keyword(s):  
Logistic Regression ◽  
Measurement Error ◽  
Regression Models ◽  
Predictor Variables ◽  
Logistic Regression Models

scholarly journals Risk Assessment Models for Wheat Fusarium Head Blight Epidemics Based on Within-Season Weather Data

2003 ◽  
Vol 93 (4) ◽  
pp. 428-435 ◽  
Author(s):  
E. D. De Wolf ◽  
L. V. Madden ◽  
P. E. Lipps
Keyword(s):  
Logistic Regression ◽  
Relative Humidity ◽  
Fusarium Head Blight ◽  
Prediction Accuracy ◽  
Regression Models ◽  
Weather Conditions ◽  
Predictor Variables ◽  
Stepwise Logistic Regression ◽  
Head Blight ◽  
Logistic Regression Models

Logistic regression models for wheat Fusarium head blight were developed using information collected at 50 location-years, including four states, representing three different U.S. wheat-production regions. Non-parametric correlation analysis and stepwise logistic regression analysis identified combinations of temperature, relative humidity, and rainfall or durations of specified weather conditions, for 7 days prior to anthesis, and 10 days beginning at crop anthesis, as potential predictor variables. Prediction accuracy of developed logistic regression models ranged from 62 to 85%. Models suitable for application as a disease warning system were identified based on model prediction accuracy, sensitivity, specificity, and availability of weather variables at crop anthesis. Four of the identified models correctly classified 84% of the 50 location-years. A fifth model that used only pre-anthesis weather conditions correctly classified 70% of the location-years. The most useful predictor variables were the duration (h) of precipitation 7 days prior to anthesis, duration (h) that temperature was between 15 and 30°C 7 days prior to anthesis, and the duration (h) that temperature was between 15 and 30°C and relative humidity was greater than or equal to 90%. When model performance was evaluated with an independent validation set (n = 9), prediction accuracy was only 6% lower than the accuracy for the original data sets. These results indicate that narrow time periods around crop anthesis can be used to predict Fusarium head blight epidemics.

Predicting Amputation and Multiple Debridements in Tooth Knuckle Injuries

2019 ◽  
Vol 24 (01) ◽  
pp. 6-12
Author(s):  
H.R. Smith ◽  
C. Conyard ◽  
J. Loveridge ◽  
R. Gunnarsson
Keyword(s):  
Logistic Regression ◽  
Cohort Study ◽  
Predictor Variables ◽  
Delayed Presentation ◽  
Multivariate Logistic Regression ◽  
Outcome Measurements ◽  
Major Predictor ◽  
Laboratory Variables ◽  
The Relationship ◽  
Historic Cohort

Background: Tooth knuckle injuries can be expensive to treat and may necessitate amputation in some cases. Several limitations exist in the literature regarding our knowledge around the factors predicting amputation and the need for multiple debridements in treating this injury.Methods: A historic cohort study of 321 patients treated for tooth knuckle injuries was undertaken. Twenty-one demographic, clinical and laboratory variables were collected. Two outcome measurements were collected - the need for amputation and the need for more than one surgical debridement. A multivariate logistic regression was performed to determine the relationship between the predictor variables and the outcome measurements.Results: Of the 321 patients examined, 1.6% required amputations and 25% required multiple debridements. Osteomyelitis was found to be a major predictor for amputation in these patients (OR = 35). Delayed presentation (OR = 1.1) and diabetes (OR = 2.6) were found to significantly increase the risk of requiring multiple debridements.Conclusions: Our models were able to predict what patients were at the greatest risk for amputation and multiple debridement. Reducing rates of osteomyelitis and delays in presentation may help reduce the incidence of amputation and reoperation in this injury.

VR Service Patterns and Employment Outcomes of Transition-Aged Youth

2016 ◽  
Vol 47 (2) ◽  
pp. 20-26
Author(s):  
Gina Oswald
Keyword(s):  
Logistic Regression ◽  
Vocational Rehabilitation ◽  
Service Provision ◽  
Employment Outcomes ◽  
Predictor Variables ◽  
Competitive Employment ◽  
Service Patterns ◽  
Successful Employment ◽  
Managerial Positions

The purpose of this study was to descriptively explore the service provision of transition-aged youth in a state vocational rehabilitation (VR) agency and to determine if predictor variables could be identified for successful employment outcomes through logistic regression. At closure, more than half the participants were closed successfully in competitive employment. The majority were working in service, clerical and sales, or professional/technical/ managerial positions after receiving VR services focused on understanding the consumer's needs and creating appropriate plans, preparing for a job, obtaining a job and then retaining employment. Implications for transition and rehabilitation practice include the necessity o[specific transition-related training for VR counselors.

Migraine Days Decline with Duration of Illness in Adolescents with Transformed Migraine

Cephalalgia ◽  
2005 ◽  
Vol 25 (7) ◽  
pp. 482-487 ◽  
Author(s):  
ME Bigal ◽  
FD Sheftell ◽  
SJ Tepper ◽  
AM Rapoport ◽  
RB Lipton
Keyword(s):  
Logistic Regression ◽  
Medication Overuse ◽  
Past History ◽  
International Headache Society ◽  
Transformed Migraine ◽  
Recent Onset ◽  
Headache Centre ◽  
Dichotomous Outcome ◽  
Long Duration ◽  
Adolescent Sample

The aim of this study was to assess the proportion of subjects with transformed migraine (TM) who have 15 or more migraine days per month as a function of duration of chronic daily headache (CDH) in an adolescent sample. CDH is a syndrome characterized by 15 or more headache days per month. In specialty care, TM is the most common type of CDH. Most adults who meet criteria for TM do not meet the International Headache Society (IHS) criteria for chronic migraine (CM). TM criteria require 15 or more headache days per month (not necessarily migraine), with a current or past history of migraine. CM requires 15 or more migraine days per month. As TM develops, attack frequency increases and the number of migraine features diminishes. If this observation is correct, individuals who meet criteria for TM but not CM may be at a later stage in the evolution of the disease, compared with those who meet criteria for CM. We reviewed charts of 267 adolescents (13-17 years) seen in a headache centre, to identify 117 with TM. We divide subjects with TM into those with recent onset (1 year) vs. longer duration (>1 year) and examined the number of migraine days per month and demographic features. We modelled predictors of CM (>15 migraine days per month) using logistic regression. Of 117 adolescents with TM, 55 (47%) had recent-onset (<1 year) and 62 (53%) had long-duration TM. Those with recent-onset TM were much more likely also to meet criteria for CM (74.5% vs. 25.8%, P < 0.001). This was verified in the TM with medication overuse subgroup (recent onset 66.7%, vs. long duration 37%, P = 0.01) and in the TM without medication overuse subgroup (62.2% vs. 19.2%, P = 0.001). Modelling the dichotomous outcome of CM (>15 days of migraine/month) in logistic regression, CM was predicted by recent onset of CDH, recent onset of migraine (<36 months), and younger ages (15 years), but not gender or use of migraine preventive drugs or medication overuse. Among adolescents with TM, CM is more likely in individuals who are young, whose episodic headache began recently, and with CDH of recent onset. These findings suggest that early in the process of transformation, migraine is more frequent, and that as CDH evolves, fewer typical attacks of IHS migraine occur.

Sign in / Sign up

Export Citation Format

Share Document