ECG lead-wire interchanges involving the right leg (RL) are not always detected. These RL lead-wire interchanges cannot be simulated in the same way as other lead-wire interchanges making database collection a necessi...
详细信息
ISBN:
(纸本)9781538645550;9781538666302
ECG lead-wire interchanges involving the right leg (RL) are not always detected. These RL lead-wire interchanges cannot be simulated in the same way as other lead-wire interchanges making database collection a necessity for algorithm development. Adult 12-lead ECGs from a single teaching hospital taken between January 2008 and July 2012 were reviewed for lead-wire interchanges by an expert electrocardiographer. Lead-wire interchanges were confirmed by comparison of serial ECGs. Positive interchanges included left arm / right leg (flat lead III, n = 134) and right arm / right leg (flat lead II, n = 139). A RL lead-wire interchange algorithm was developed by bootstrap aggregation of decision trees with 5-fold cross validation. Test results were summed over the 5-fold cross validation on the partitions not used for training. ECG features included maximum and minimum QRS and T-wave voltages for ECG leads I, II and III. The Haisty algorithm for RA-RL interchange was tested for comparison. algorithm performance was quantified by sensitivity (SE), specificity (SP) and estimated positive predictive value (PPV) based on SE, SP and realistic prevalence. For a prevalence of 0.2%, performance in SE, SP and PPV was: Haisty, 94, 99.4, 24;tree RA-RL: 84, 99.9, 57;tree LA-RL: 87, 99.9, and 57%. Even though SP was high for all three algorithms, the estimated PPVs were modest due to the low prevalence. Conclusion: Lead-wire interchanges involving the right leg wire can be detected with good sensitivity and high specificity. The higher specificity of the tree based algorithms results in more than twice the PPV of the Haisty algorithm.
The paper uses machine learning methods to deal with the problem of reducing the cost of applying mutation testing. A method of classifying mutants of a program using structural similarity is proposed. To calculate su...
详细信息
ISBN:
(纸本)9788394625375
The paper uses machine learning methods to deal with the problem of reducing the cost of applying mutation testing. A method of classifying mutants of a program using structural similarity is proposed. To calculate such a similarity each mutant is firstly converted into a hierarchical graph, which represents the mutant's control flow, variables and conditions. Then using such a graph form graph kernels are introduced to calculate similarity among mutants. The classification algorithm is then applied for prediction. This approach helps to lower the number of mutants which have to be executed. An experimental validation of this approach is also presented.
In order to reduce software energy consumption, a lot of studies have been carried out focusing on the difference of implementation, such as API and algorithm. However, we hypothesize that there is a strong correlatio...
详细信息
ISBN:
(纸本)9781538603673
In order to reduce software energy consumption, a lot of studies have been carried out focusing on the difference of implementation, such as API and algorithm. However, we hypothesize that there is a strong correlation between total energy consumption of a program and duration of its execution. If this hypothesis is correct, reducing energy consumption is equal to decreasing duration. Experimental results reveal that there is a strong positive correlation between them, and its correlation coefficient is higher than 0.9. We also find that memory usage is weakly correlated with total energy consumption. As a result, we conclude that if developers want to reduce software energy consumption, they should firstly decrease duration of execution, and secondly reduce memory usage.
We present a combined method of classical signal analysis and machine learning algorithms for the automated classification of 1-lead ECG recordings, which was developed in the course of the Computing in Cardiology Cha...
详细信息
ISBN:
(纸本)9781538645550;9781538666302
We present a combined method of classical signal analysis and machine learning algorithms for the automated classification of 1-lead ECG recordings, which was developed in the course of the Computing in Cardiology Challenge 2017. To classify ECG recordings into the four classes as defined for the Challenge (normal, suspicious to AF, suspicious to other arrhythmia, noise) we used MATLABand a set of algorithms for detection of beats, wave point detection on detected beats, quality evaluation of the detection, averaging of beats, beat classification, rhythm classification and many more. We extracted a variety of features from both time and frequency domain etc. as input features for the classifier. A total of 380 features were used to train a Random Forest - based classifier (bagged decision trees). Since classes for the Challenge were severely unbalanced, weights based on the class distribution were applied. To train the classifier and for our internal evaluation we used cross-validation on all available ECGS from the training-set. 10-fold cross-validated F-1 score on the training set is 0.83. Final F-1 score from the official challenge evaluation on the enhanced dataset is 0.81, which is quite close to the other top performing algorithms.
The development of new technology such as wearables that record high-quality single channel ECG, provides an opportunity for ECG screening in a larger population, especially for atrial fibrillation screening. The main...
详细信息
ISBN:
(纸本)9781538645550;9781538666302
The development of new technology such as wearables that record high-quality single channel ECG, provides an opportunity for ECG screening in a larger population, especially for atrial fibrillation screening. The main goal of this study is to develop an automatic classification algorithm for normal sinus rhythm (NSR), atrial fibrillation (AF), other rhythms (O), and noise from a single channel short ECG segment (9-60 seconds). For this purpose, signal quality index (SQI) along with dense convolutional neural networks was used. Two convolutional neural network (CNN) models (main model that accepts 15 seconds ECG and secondary model that processes 9 seconds shorter ECG) were trained using the training data set. If the recording is determined to be of low quality by SQI, it is immediately classified as noisy. Otherwise, it is transformed to a time-frequency representation and classified with the CNN as NSR, AF, O, or noise. At the final step, a feature-based post-processing algorithm classifies the rhythm as either NSR or O in case the CNN model's discrimination between the two is indeterminate. The best result achieved at the official phase of the PhysioNet/CinC challenge on the blind test set was 0.80 (F1 for NSR, AF, and O were 0.90, 0.80, and 0.70, respectively).
Significant contrast in visible wavelength Mueller matrix images for healthy and pre-cancerous regions of excised cervical tissue is shown. A novel classification algorithm is used to compute a test statistic from a s...
详细信息
ISBN:
(纸本)9781510612815;9781510612808
Significant contrast in visible wavelength Mueller matrix images for healthy and pre-cancerous regions of excised cervical tissue is shown. A novel classification algorithm is used to compute a test statistic from a small patient population.
Classification of Atrial Fibrillation from diverse electro-cardiographic (ECG) signals is the challenging objective of the 2017 Physionet Challenge. We suggest a Long Short Term Memory (LSTM) network, which learns pat...
详细信息
ISBN:
(纸本)9781538645550;9781538666302
Classification of Atrial Fibrillation from diverse electro-cardiographic (ECG) signals is the challenging objective of the 2017 Physionet Challenge. We suggest a Long Short Term Memory (LSTM) network, which learns patterns directly from pre-computed QRS complex features that classifies ECG signals. Although our architecture is considered deep, it only consists of 1791 parameters. The result is an accurate, lightweight solution that classifies ECG records as Normal, Atrial fibrillation, Other or Too noisy with final challenge score of 0.78.
Cardiac arrhythmias are the leading cause of death in the western world, where atrial fibrillation (AF) is the most common arrhythmias. The PhysioNet/CinC 2017 Challenge aimed to trigger a design of an algorithm that ...
详细信息
ISBN:
(纸本)9781538645550;9781538666302
Cardiac arrhythmias are the leading cause of death in the western world, where atrial fibrillation (AF) is the most common arrhythmias. The PhysioNet/CinC 2017 Challenge aimed to trigger a design of an algorithm that accurately classifies short single ECG lead record to 4 categories: normal rhythm, atrial fibrillation, noisy segment or other arrhythmias. The algorithm was optimized on randomly selected records out of the challenge learning set (8528 records after reassuring it includes 60.43% of normal records, 0.54% of noisy records, 9.04% of AF records and 30% of other rhythm disturbance) and tested on hidden test database. A novel R peak detector was used to accurately detect the R peaks. Based on the R peak annotation, the P, Q, S and T peaks were detected and ECG beat morphology was extracted. Quadratic SVM classifier that include combination of 62 features was used to classify the short ECG record to one of the four categories mentioned above. For records which were classified as "normal" additional neural network classifier was applied. Our algorithm reached results of total score (F1) of 0.8 (ranked 24 out a total of 90 open-source software entries), whereas normal rhythm score (F-1n) was 0.9, AF rhythm score (F-1a) of 0.81, and other rhythm score (F-1o) of 0.69.
An improved KNN text classification algorithm based on Simhash has been proposed by introducing Simhash and the average Hamming distance of adjacent texts as a unit, which solves the problems caused by data imbalance ...
详细信息
ISBN:
(纸本)9781538607718
An improved KNN text classification algorithm based on Simhash has been proposed by introducing Simhash and the average Hamming distance of adjacent texts as a unit, which solves the problems caused by data imbalance and the large computational overhead in the traditional KNN text classification algorithms. Experimental results demonstrate that the proposed algorithm performs a higher precision, a higher recall and a better F1 value, which shows the validity of the proposed algorithm.
The PhysioNet/Computing in Cardiology (CinC) Challenge 2017 focused on differentiating AF from noise, normal or other rhythms in short term (from 9-61 s) ECG recordings performed by patients. A total of 12,186 ECGs we...
详细信息
ISBN:
(纸本)9781538645550;9781538666302
The PhysioNet/Computing in Cardiology (CinC) Challenge 2017 focused on differentiating AF from noise, normal or other rhythms in short term (from 9-61 s) ECG recordings performed by patients. A total of 12,186 ECGs were used: 8,528 in the public training set and 3,658 in the private hidden test set. Due to the high degree of inter-expert disagreement between a significant fraction of the expert labels we implemented a mid-competition bootstrap approach to expert relabeling of the data, levering the best performing Challenge entrants' algorithms to identify contentious labels. A total of 75 independent teams entered the Challenge using a variety of traditional and novel methods, ranging from random forests to a deep learning approach applied to the raw data in the spectral domain. Four teams won the Challenge with an equal high F1 score (averaged across all classes) of 0.83, although the top 11 algorithms scored within 2% of this. A combination of 45 algorithms identified using LASSO achieved an F1 of 0.87, indicating that a voting approach can boost performance.
暂无评论