Understanding sensitivity, specificity and predictive values

Stojanović, Miodrag; Andjelkovic-Apostolovic, Marija; Stojanović, Dijana; Milošević, Zoran; Ignjatović, Aleksandra; Mitic-Lakusic, Vesna; Golubović, Mlađjan

doi:10.2298/vsp1411062s

Cited by 50 publications

(41 citation statements)

References 13 publications

(8 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This is valuable for the classifications because it is preferable to inaccurately assume an urge is present (and risk unnecessary urge query) than miss a real urge (and risk not intervening to prevent smoking). In addition, sensitivity and specificity are inversely proportional [47,48]. That can also explain why the specificity is low: as the sensitivity increases, the specificity decreases and vice versa.…”

Section: Discussionmentioning

confidence: 99%

Classifying smoking urges via machine learning

Dumortier

Beckjord

Shiffman

et al. 2016

Computer Methods and Programs in Biomedicine

View full text Add to dashboard Cite

Background and objective Smoking is the largest preventable cause of death and diseases in the developed world, and advances in modern electronics and machine learning can help us deliver real-time intervention to smokers in novel ways. In this paper, we examine different machine learning approaches to use situational features associated with having or not having urges to smoke during a quit attempt in order to accurately classify high-urge states. Methods To test our machine learning approaches, specifically, Bayes, discriminant analysis and decision tree learning methods, we used a dataset collected from over 300 participants who had initiated a quit attempt. The three classification approaches are evaluated observing sensitivity, specificity, accuracy and precision. Results The outcome of the analysis showed that algorithms based on feature selection make it possible to obtain high classification rates with only a few features selected from the entire dataset. The classification tree method outperformed the naive Bayes and discriminant analysis methods, with an accuracy of the classifications up to 86%. These numbers suggest that machine learning may be a suitable approach to deal with smoking cessation matters, and to predict smoking urges, outlining a potential use for mobile health applications. Conclusions In conclusion, machine learning classifiers can help identify smoking situations, and the search for the best features and classifier parameters significantly improves the algorithms’ performance. In addition, this study also supports the usefulness of new technologies in improving the effect of smoking cessation interventions, the management of time and patients by therapists, and thus the optimization of available health care resources. Future studies should focus on providing more adaptive and personalized support to people who really need it, in a minimum amount of time by developing novel expert systems capable of delivering real-time interventions.

show abstract

Section: Discussionmentioning

confidence: 99%

Classifying smoking urges via machine learning

Dumortier

Beckjord

Shiffman

et al. 2016

Computer Methods and Programs in Biomedicine

View full text Add to dashboard Cite

show abstract

“…Standard analyses of the CABS sensitivity, specificity, positive predictive value, and negative predictive value were conducted to evaluate the instrument's potential utility as a screening tool (Katz, ; Stojanovic et al, ; Table ). Estimates of the area under the receiver operating characteristic (ROC) curve were used to assess the predictive performance of the CABS instrument, after collapsing CABS scores into two categories: bullied and not bullied.…”

Section: Methodsmentioning

confidence: 99%

“…The construct validity of the CABS screen was evaluated using the known-groups method (Burns & Grove, 2009). Mean CABS scores for subjects classified as having and not having experienced bullying based on the CBVS and the R- Standard analyses of the CABS sensitivity, specificity, positive predictive value, and negative predictive value were conducted to evaluate the instrument's potential utility as a screening tool (Katz, 2006;Stojanovic et al, 2014; Table 2). Estimates of the area under the receiver operating characteristic (ROC) curve were used to assess the predictive performance of the CABS instrument, after collapsing CABS scores into two categories: bullied and not bullied.…”

Section: Instrument Functioning In Comparison To Existing Standardsmentioning

confidence: 99%

“…Potential cut-points for the CABS (potential scores that would determine a positive CABS screen) were explored through visual Positive Predictive Value (Stojanovic et al, 2014) The probability that a participant with a positive CABS screen has actually been exposed to bullying [Participants with true positive results the total number of participants with positive results] x 100%…”

Section: Instrument Functioning In Comparison To Existing Standardsmentioning

confidence: 99%

“…Negative Predictive Value (Stojanovic et al, 2014) The probability that a participant with a negative CABS screen has not been exposed to bullying [Participants with a true negative result the total number participants with a negative result] x 100% Accuracy (Katz, 2006) The proportion of correct bullying exposure "diagnoses" made by the CABS screen [true positives + true negatives] total sample size Likelihood Ratio for a Positive Test (Katz, 2006) The ratio of the probability of a positive CABS screen in participants exposed to bullying to the probability of a positive CABS screen in a participant without bullying exposure Sensitivity ÷ (1-specificity)…”

Section: Instrument Functioning In Comparison To Existing Standardsmentioning

confidence: 99%

See 2 more Smart Citations

The Child Adolescent Bullying Scale (CABS): Psychometric evaluation of a new measure

Strout

Vessey

DiFazio

et al. 2018

Research in Nursing & Health

View full text Add to dashboard Cite

While youth bullying is a significant public health problem, healthcare providers have been limited in their ability to identify bullied youths due to the lack of a reliable, and valid instrument appropriate for use in clinical settings. We conducted a multisite study to evaluate the psychometric properties of a new 22-item instrument for assessing youths' experiences of being bullied, the Child Adolescent Bullying Scale (CABS). The 20 items summed to produce the measure's score were evaluated here. Diagnostic performance was assessed through evaluation of sensitivity, specificity, predictive values, and area under receiver operating characteristic (AUROC) curve. A sample of 352 youths from diverse racial, ethnic, and geographic backgrounds (188 female, 159 male, 5 transgender, sample mean age 13.5 years) were recruited from two clinical sites. Participants completed the CABS and existing youth bullying measures. Analyses grounded in classical test theory, including assessments of reliability and validity, item analyses, and principal components analysis, were conducted. The diagnostic performance and test characteristics of the CABS were also evaluated. The CABS is comprised of one component, accounting for 67% of observed variance. Analyses established evidence of internal consistency reliability (Cronbach's α = 0.97), construct and convergent validity. Sensitivity was 84%, specificity was 65%, and the AUROC curve was 0.74 (95% CI: 0.69-0.80). Findings suggest that the CABS holds promise as a reliable, valid tool for healthcare provider use in screening for bullying exposure in the clinical setting.

show abstract

Iterative magnitude pruning-based light-version of AlexNet for skin cancer classification

Medhat,

Abdel-Galil,

Aboutabl

et al. 2023

Neural Comput & Applic

View full text Add to dashboard Cite

Convolutional Neural Networks (CNN) with different architectures have shown promising results in skin cancer diagnosis. However, CNN has a high computational cost, which makes the need for a light version of CNN a desirable step. This version can be used on small devices, such as mobile phones or tablets. A light version can be created using pruning techniques. In this study, iterative magnitude pruning (IMP) is utilized. This method depends on pruning the network iteratively. The IMP method is applied on AlexNet with transfer learning (TL) and data augmentation. The proposed IMP AlexNet with TL is applied on three different skin cancer datasets which are PAD-UFES-20, MED-NODE, and PH2 dataset. The datasets used are a combination of smartphone, dermoscopic, and non-dermoscopic images. Different CNN versions are applied on the same datasets for comparison with IMP AlexNet. The CNNs used are VGG-16, ShuffleNet, SqueezNet, DarkNet-19, DarkNet-53, and Inception-v3. The proposed IMP AlexNet achieved accuracies of 97.62%, 96.79%, and 96.75%, with accuracy losses of 1.53%, 2.3%, and 2.2%, respectively, compared to the original AlexNet. In addition, the proposed IMP AlexNet requires less running time and memory usage than the traditional AlexNet. The average running time for IMP AlexNet is 0.45 min, 0.28 min, and 0.3 min, for PAD-UFES-20, MED-NODE, and PH2 datasets, respectively. The average RAM usage with IMP AlexNet is 1.8 GB, 1.6 GB, and 1.7 GB, respectively. IMP AlexNet accelerates the average running time by approximately 15 times that of the traditional AlexNet and reduces the average RAM used by 40%.

show abstract

Understanding sensitivity, specificity and predictive values

Abstract: Key words: data interpretation, statistical; sensitivity and specificity; predictive value of tests.Klju ne re i: statisti ka analiza podataka; osetljivost i specifi nost; testovi, prognosti ka vrednost.

Cited by 50 publications

References 13 publications

Classifying smoking urges via machine learning

Classifying smoking urges via machine learning

The Child Adolescent Bullying Scale (CABS): Psychometric evaluation of a new measure

Iterative magnitude pruning-based light-version of AlexNet for skin cancer classification

Contact Info

Product

Resources

About