General method for automated feature extraction and selection and its application for gender classification and biomechanical knowledge discovery of sex differences in spinal posture during stance and gait

Dindorf, Carlo; Konradi, Jürgen; Wolf, Claudia; Taetz, Bertram; Bleser, Gabriele; Huthwelker, Janine; Drees, Philipp; Fröhlich, Michael; Betz, Ulrich

doi:10.1080/10255842.2020.1828375

Cited by 14 publications

(18 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The training data for the respective pathology should, therefore, be expanded, especially for subjects that show similar characteristics to the misclassified subject, so that the model is able to map the respective characteristics during the training phase. Further feature engineering through an automated feature extraction [ 37 ] or the inclusion of global spinal parameters (e.g., lordosis and kyphosis angle) might also be promising. For the OCSVM, a further possible reason for misclassification was that hyperparameters were not optimally chosen due to the hyperparameter search, because it was only based on the validation set performance related to the healthy subjects.…”

Section: Discussionmentioning

confidence: 99%

“…The method allowed to measure the spine in all body planes without the use of invasive or radiation-based approaches or extensive preparation. Recently, in addition to static measurements, this method has proven useful in measuring dynamic spinal data [ 37 , 38 ]. (1 The dataset is part of the dissertation project of Friederike Werthmann; 2 The dataset is part of the dissertation project of Claudia Wolf).…”

Section: Methodsmentioning

confidence: 99%

“…Fifty-five static parameters provided by the system were used (pelvic obliquity (°), pelvic torsion (dimples) (°), pelvic inclination (dimples) (°), pelvic rotation (°), and orientation of VP, T1–12, and L1–L4 in all planes (°)) for modeling. For a detailed description of the parameters, see [ 37 , 39 ] and Appendix A . Data of healthy subjects were used for training the OCSVM.…”

Section: Methodsmentioning

confidence: 99%

See 2 more Smart Citations

Classification and Automated Interpretation of Spinal Posture Data Using a Pathology-Independent Classifier and Explainable Artificial Intelligence (XAI)

Dindorf

Konradi

Wolf

et al. 2021

Sensors

Self Cite

View full text Add to dashboard Cite

Clinical classification models are mostly pathology-dependent and, thus, are only able to detect pathologies they have been trained for. Research is needed regarding pathology-independent classifiers and their interpretation. Hence, our aim is to develop a pathology-independent classifier that provides prediction probabilities and explanations of the classification decisions. Spinal posture data of healthy subjects and various pathologies (back pain, spinal fusion, osteoarthritis), as well as synthetic data, were used for modeling. A one-class support vector machine was used as a pathology-independent classifier. The outputs were transformed into a probability distribution according to Platt’s method. Interpretation was performed using the explainable artificial intelligence tool Local Interpretable Model-Agnostic Explanations. The results were compared with those obtained by commonly used binary classification approaches. The best classification results were obtained for subjects with a spinal fusion. Subjects with back pain were especially challenging to distinguish from the healthy reference group. The proposed method proved useful for the interpretation of the predictions. No clear inferiority of the proposed approach compared to commonly used binary classifiers was demonstrated. The application of dynamic spinal data seems important for future works. The proposed approach could be useful to provide an objective orientation and to individually adapt and monitor therapy measures pre- and post-operatively.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Classification and Automated Interpretation of Spinal Posture Data Using a Pathology-Independent Classifier and Explainable Artificial Intelligence (XAI)

Dindorf

Konradi

Wolf

et al. 2021

Sensors

Self Cite

View full text Add to dashboard Cite

show abstract

“…This factor quantifies the level of impurity/inequality of the samples assigned to each tree node based on the split done with respect to its parent. The features were sorted based on their Gini importance in a descending order in order to explicitly eliminate unimportant features and select a predefined number of features with higher significance (N s ) as the input variables for other classifiers, as successfully performed in numerous studies [45,46,[48][49][50][51][52][53].…”

Section: Curse Of Dimensionality and Necessity Of Dimensionality Reductionmentioning

confidence: 99%

Frailty Level Classification of the Community Elderly Using Microsoft Kinect-Based Skeleton Pose: A Machine Learning Approach

Akbari

Nikkhoo

Wang

et al. 2021

Sensors

View full text Add to dashboard Cite

Frailty is one of the most important geriatric syndromes, which can be associated with increased risk for incident disability and hospitalization. Developing a real-time classification model of elderly frailty level could be beneficial for designing a clinical predictive assessment tool. Hence, the objective of this study was to predict the elderly frailty level utilizing the machine learning approach on skeleton data acquired from a Kinect sensor. Seven hundred and eighty-seven community elderly were recruited in this study. The Kinect data were acquired from the elderly performing different functional assessment exercises including: (1) 30-s arm curl; (2) 30-s chair sit-to-stand; (3) 2-min step; and (4) gait analysis tests. The proposed methodology was successfully validated by gender classification with accuracies up to 84 percent. Regarding frailty level evaluation and prediction, the results indicated that support vector classifier (SVC) and multi-layer perceptron (MLP) are the most successful estimators in prediction of the Fried’s frailty level with median accuracies up to 97.5 percent. The high level of accuracy achieved with the proposed methodology indicates that ML modeling can identify the risk of frailty in elderly individuals based on evaluating the real-time skeletal movements using the Kinect sensor.

show abstract

“…In addition to this producing machine-understandable and processable data from existing resources (rather than to produce data from scratch) is one of the key challenges for computer science (Koumenides et al, 2010;Hochtl, Reichst & Adter, 2011), (especially in terms of information retrieval (Manning, Raghavan & Schtze, 2008) and feature extraction Dindorf et al, 2020). This topic is becoming more and more important because of its use in creating smart applications and mashups for data acquisition and processing to compute behaviors in knowledge society (DiFranzo et al, 2011).…”

Section: Introductionmentioning

confidence: 99%

LOPDF: a framework for extracting and producing open data of scientific documents for smart digital libraries

Aslam

2021

PeerJ Computer Science

View full text Add to dashboard Cite

Background Results of scientific experiments and research work, either conducted by individuals or organizations, are published and shared with scientific community in different types of scientific publications such as books, chapters, journals, articles, reference works and reference works entries. One aspect of these documents is their contents and the other is metadata. Metadata of scientific documents could be used to increase mutual cooperation, find people with common interest and research work, and to find scientific documents in the matching domains. The major issue in getting these benefits from metadata of scientific publications is availability of these data in unstructured (or semi-structured) format so that it can not be used to ask smart queries that can help in computing and performing different types of analysis on scientific publications data. Also, acquisition and smart processing of publications data is a complicated as well as time and resource consuming task. Methods To address this problem we have developed a generic framework named as Linked Open Publications Data Framework (LOPDF). The LOPDF framework can be used to crawl, process, extract and produce machine understandable data (i.e., LOD) about scientific publications from different publisher specific sources such as portals, XML export and websites. In this paper we present the architecture, process and algorithm that we developed to process textual publications data and to produce semantically enriched data as RDF datasets (i.e., open data). Results The resulting datasets can be used to make smart queries by making use of SPARQL protocol. We also present the quantitative as well as qualitative analysis of our resulting datasets which ultimately can be used to compute the research behavior of organizations in rapidly growing knowledge society. Finally, we present the potential usage of producing and processing such open data of scientific publications and how results of performing smart queries on resulting open datasets can be used to compute the impact and perform different types of analysis on scientific publications data.

show abstract

General method for automated feature extraction and selection and its application for gender classification and biomechanical knowledge discovery of sex differences in spinal posture during stance and gait

Cited by 14 publications

References 45 publications

Classification and Automated Interpretation of Spinal Posture Data Using a Pathology-Independent Classifier and Explainable Artificial Intelligence (XAI)

Classification and Automated Interpretation of Spinal Posture Data Using a Pathology-Independent Classifier and Explainable Artificial Intelligence (XAI)

Frailty Level Classification of the Community Elderly Using Microsoft Kinect-Based Skeleton Pose: A Machine Learning Approach

LOPDF: a framework for extracting and producing open data of scientific documents for smart digital libraries

Contact Info

Product

Resources

About