1 September 2020 in Journal of the American Medical Informatics Association
Objective: We sought to predict if patients with type 2 diabetes mellitus (DM2) would develop 10 selected complications. Accurate prediction of complications could help with more targeted measures that would prevent or slow down their development.Materials and Methods: Experiments were conducted on the Healthcare Cost and Utilization Project State Inpatient Databases of California for the period of 2003 to 2011. Recurrent neural network (RNN) long short-term memory (LSTM) and RNN gated recurrent unit (GRU) deep learning methods were designed and compared with random forest and multilayer perceptron traditional models. Prediction accuracy of selected complications were compared on 3 settings corresponding to minimum number of hospitalizations between diabetes diagnosis and the diagnosis of complications.Results: The diagnosis domain was used for experiments. The best results were achieved with RNN GRU model, followed by RNN LSTM model. The prediction accuracy achieved with RNN GRU model was between 73% (myocardial infarction) and 83% (chronic ischemic heart disease), while accuracy of traditional models was between 66% – 76%.Discussion: The number of hospitalizations was an important factor for the prediction accuracy. Experiments with 4 hospitalizations achieved significantly better accuracy than with 2 hospitalizations. To achieve improved accuracy deep learning models required training on at least 1000 patients and accuracy significantly dropped if training datasets contained 500 patients. The prediction accuracy of complications decreases over time period. Considering individual complications, the best accuracy was achieved on depressive disorder and chronic ischemic heart disease.Conclusions: The RNN GRU model was the best choice for electronic medical record type of data, based on the achieved results.
24 September 2021 in IEEE Transactions on Instrumentation and Measurement
Event classification is one of the central components of automated disturbance analysis based on phasor measurement unit (PMU) measurements. Obtaining high-quality event labels remains a challenge for supervised learning-based classification of local and system-wide events in power grids due to its labor-intensive requirement. We present a sensitivity study considering rapidly refined, partially, and fully inspected event labels that leads to evidence that hierarchical convolutional neural networks (HCNNs) outperform traditional classification models regardless of the quality of the available event labels. It is demonstrated that performance similar to the one obtained using entirely domain-driven labeling can be achieved as long as the involved expert does not mislabel more than ~5% of the event data captured by PMU measurements.
10 September 2021 in IEEE Access
Event detection in electrical grids is a challenging problem for machine learning methods due to spatiotemporally nonstationary systems and the inability to automate event labeling in high-volume data such as PMU measurements. As a result, the existing historical event logs created manually do not correlate well with the corresponding PMU measurements due to scarce and temporally imprecise labels. Trying to overcome this problem by extending event logs to a complete set of labeled events is very costly and often infeasible. We focused on utilizing a transfer learning model to reduce the need for additional data labeling by leveraging some labeled data instances available from a small number of well-defined event detection task. To demonstrate the feasibility, we tested our approach on a large dataset collected by 38 PMUs from the Western Interconnection of the U.S.A. over two years. The model evaluation performed based on varying percentages of labeled source data corresponding to ~20-700 characteristic events on different sizes of time windows ranging from 2-seconds to 1-minute demonstrates that the developed method can significantly improve automated event detection based on PMU measurements when extensive labeling is costly or impossible to obtain. When compared to the state-of-the-art machine learning algorithms (unsupervised, semi-supervised, and supervised), the results show that the transfer learning method has significantly better performances when detecting events by learning from as low as 20 representative labeled data instances.
24 March 2022 in Informatics in Medicine Unlocked