Threshold optimization for F measure of macro-averaged precision and recall

Berger, Anna; Гуда, С. А.

doi:10.1016/j.patcog.2020.107250

Cited by 31 publications

(9 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For each identified weld region, the identification score of the weld region is expressed as a confidence level. e precision-recall curves are plotted [26].…”

Section: Evaluation Methods For the Performance Of Weld Areamentioning

confidence: 99%

Machine Vision-Based Object Detection Strategy for Weld Area

Liu

Chen

Huang

2022

Scientific Programming

View full text Add to dashboard Cite

For the noisy industrial environment, the welded parts will have different types of defects in the weld area during the welding process, which need to be polished, and there are disadvantages such as low efficiency and high labor intensity when polishing manually; machine vision is used to automate the polishing and achieve continuous and efficient work. In this study, the Faster R-CNN object detection algorithm of two-stage is used to investigate the relationship between flops and the number of network parameters on the model by using a V-shaped welded thick plate as the research object and establishing the workpiece dataset with different lighting and angles, using six regional candidate networks for migration learning, comparing the convergence degree of different Batch and Mini-Batch on the model, and exploring the relationship between flops and the number of network parameters on the model. The optimal learning rate is selected for training to form a weld area object detection network based on the weld plate workpiece under few samples. The study shows that the VGG16 model is the best in weld seam area recognition with 91.68% average accuracy and 25.02 ms average detection time in the validation set, which can effectively identify weld seam areas in various industrial environments and provide location information for subsequent automatic grinding of robotic arms.

show abstract

“…For each identified weld region, the identification score of the weld region is expressed as a confidence level. e precision-recall curves are plotted [26].…”

Section: Evaluation Methods For the Performance Of Weld Areamentioning

confidence: 99%

Machine Vision-Based Object Detection Strategy for Weld Area

Liu

Chen

Huang

2022

Scientific Programming

View full text Add to dashboard Cite

show abstract

“…The procedure was repeated two times. Subsequently, we examined two estimation criteria: macro-averaged (MA) and weighted average (WA) values [46][47][48], with four evaluation metrics [49] (balanced accuracy, precision, recall, and F1-measure), which were calculated from one confusion matrix accumulated during the cross-validation. Macro-averaged and weighted average criteria were calculated using (1) and (2):…”

Section: F Performance Evaluation Metricsmentioning

confidence: 99%

Machine Learning to Identify Psychomotor Behaviors of Delirium for Patients in Long-Term Care Facility

Son

Kang

Lee

et al. 2022

IEEE J. Biomed. Health Inform.

View full text Add to dashboard Cite

This study aimed to develop accurate and explainable machine learning models for three psychomotor behaviors of delirium for hospitalized adult patients. A prospective pilot study was conducted with 33 participants admitted to a long-term care facility between August 10 and 25, 2020. During the pilot study, we collected 560 cases that included 33 clinical variables and the survey items from the short confusion assessment method (S-CAM), and developed a mobile-based application. Multiple machine learning algorithms, including four rule-mining algorithms (C4.5, CBA, MCAR, and LEM2) and four other statistical learning algorithms (LR, ANNs, SVMs with three kernel functions, and random forest), were validated by paired Wilcoxon signed-rank tests on both macro-averaged F1 and weighted average F1-measures during the 10-times stratified 2-fold crossvalidation. The LEM2 algorithm achieved the best prediction performance (macro-averaged F1-measure of 49.35%; weighted average F1-measure of 96.55%), correctly identifying adult patients at delirium risk. In the pairwise comparison between predictive powers observed from independent models, the LEM2 model showed a medium or large effect size between 0.4925 and 0.8766 when compared with LR, ANN, SVM with RBF, and MCAR models. We have confirmed that acute consciousness in S-CAM assessment is closely associated with different predictors for screening three psychomotor behaviors of delirium: 1) education level, dementia type or its level, sleep disorder, dehydration, and infection in mixed-type delirium; 2) gender, education level, dementia type, dehydration, bedsores, and foley catheter in hyperactive delirium; and 3) pain, sleep disorder, and haloperidol use in hypoactive delirium.

show abstract

“…The LSTM model has removed all irrelevant words, and the tagging is performed only on keywords; iv) The cancer name and drug list. The performance of the output is evaluated through the following metrics; micro and macro F-measure alongside recall and Int J Artif Intell ISSN: 2252-8938 Bi-directional long short term memory using recurrent neural network for … (Rashmi Siddalingappa) 97 precision scores [35]. This is because macro scores show how the system performs across all the data provided in the testing phase and micro score indicates the performance for every input.…”

Section: Evaluation Metricsmentioning

confidence: 99%

Bi-directional long short term memory using recurrent neural network for biological entity recognition

Siddalingappa

Sekar

2022

IJ-AI

View full text Add to dashboard Cite

<p>Biomedical named entity recognition (NER) aims at identifying medical entities from unstructured data. A quintessential task in the supervision of biological databases is handling biomedical terms such as cancer type, DeoxyriboNucleic and RiboNucleic Acid, gene and protein name, and others. However, due to the massive size of online medical repositories, data processing becomes a challenge for a gazetteer without proper annotation. The traditional NER systems depend on feature engineering that is tedious and time-consuming. The research study presents a new model for Bio-NER using recurrent neural network. Unlike existing approaches, the proposed method uses bidirectional traversing with GloVe vector modelling performed at character and word levels. Bio-NER is performed in three stages; firstly, the relevant medical entities in electronic medical records from PubMed were extracted using the skip-gram model. Secondly, a vector representation for each word is created through the 1-hot method. Thirdly, the weights of the recurrent neural network (RNN) layers are adjusted using backward propagation. Finally, the long-short-term memory cells store the previously encountered medical entity to tackle context-dependency. The accuracy and F-score are calculated for each medical entity type. The MacroR, MacroP, and MacroF are equal to 0.86, 0.88, and 0.87. The overall accuracy achieved was 94%.</p>

show abstract

Threshold optimization for F measure of macro-averaged precision and recall

Cited by 31 publications

References 9 publications

Machine Vision-Based Object Detection Strategy for Weld Area

Machine Vision-Based Object Detection Strategy for Weld Area

Machine Learning to Identify Psychomotor Behaviors of Delirium for Patients in Long-Term Care Facility

Bi-directional long short term memory using recurrent neural network for biological entity recognition

Contact Info

Product

Resources

About