Experimental Evaluation of Machine Learning Methods for Robust Received Signal Strength-Based Visible Light Positioning.

In this work, the use of Machine Learning methods for robust Received Signal Strength (RSS)-based Visible Light Positioning (VLP) is experimentally evaluated. The performance of Multilayer Perceptron (MLP) models and Gaussian processes (GP) is investigated when using relative RSS input features. The experimental set-up for the RSS-based VLP technology uses light-emitting diodes (LEDs) transmitting intensity modulated light and a single photodiode (PD) as a receiver. The experiments focus on achieving robustness to cope with unknown received signal strength modifications over time. Therefore, several datasets were collected, where per dataset either the LEDs transmitting power is modified or the PD aperture is partly obfuscated by dust particles. Two relative RSS schemes are investigated. The first scheme uses the maximum received light intensity to normalize the received RSS vector, while the second approach obtains RSS ratios by combining all possible unique pairs of received intensities. The Machine Learning (ML) methods are compared to a relative multilateration implementation. It is demonstrated that the adopted MLP and GP models exhibit superior performance and higher robustness when compared to the multilateration strategies. Furthermore, when comparing the investigated ML models, the GP model is proven to be more robust than the MLP for the considered scenarios.

Netter: re-ranking gene network inference predictions using structural network properties.

BACKGROUND: Many algorithms have been developed to infer the topology of gene regulatory networks from gene expression data. These methods typically produce a ranking of links between genes with associated confidence scores, after which a certain threshold is chosen to produce the inferred topology. However, the structural properties of the predicted network do not resemble those typical for a gene regulatory network, as most algorithms only take into account connections found in the data and do not include known graph properties in their inference process. This lowers the prediction accuracy of these methods, limiting their usability in practice. RESULTS: We propose a post-processing algorithm which is applicable to any confidence ranking of regulatory interactions obtained from a network inference method which can use, inter alia, graphlets and several graph-invariant properties to re-rank the links into a more accurate prediction. To demonstrate the potential of our approach, we re-rank predictions of six different state-of-the-art algorithms using three simple network properties as optimization criteria and show that Netter can improve the predictions made on both artificially generated data as well as the DREAM4 and DREAM5 benchmarks. Additionally, the DREAM5 E.coli. community prediction inferred from real expression data is further improved. Furthermore, Netter compares favorably to other post-processing algorithms and is not restricted to correlation-like predictions. Lastly, we demonstrate that the performance increase is robust for a wide range of parameter settings. Netter is available at http://bioinformatics.intec.ugent.be. CONCLUSIONS: Network inference from high-throughput data is a long-standing challenge. In this work, we present Netter, which can further refine network predictions based on a set of user-defined graph properties. Netter is a flexible system which can be applied in unison with any method producing a ranking from omics data. It can be tailored to specific prior knowledge by expert users but can also be applied in general uses cases. Concluding, we believe that Netter is an interesting second step in the network inference process to further increase the quality of prediction.

FloReMi: Flow density survival regression using minimal feature redundancy.

Advances in flow cytometry bioinformatics have resulted in a wide variety of clustering, classification and visualization techniques. To objectively evaluate the performance of such methods, common benchmarks such as the FlowCAP initiative have proven to be of great value. In this work, we report on a novel method, FloReMi, which was developed to tackle the most recent FlowCAP IV challenge. This challenge was formulated as a survival modeling problem, where participants were expected to design a model to predict the time until progression to AIDS for HIV patients. It is known that variability in progression rate cannot be fully predicted by simple CD4(+) T cell counts. However, it is hypothesized that the immunopathogenesis established early in HIV already indicates the course of future disease. Adequately estimating the progression rate of HIV patients is crucial in their treatment. Using an automated pipeline to preprocess the data, and subsequently identify and select informative cell subsets, a survival regression method based on random survival forests was built, which obtained the best results of all submitted approaches to the FlowCAP IV challenge.

A benchmark for evaluation of algorithms for identification of cellular correlates of clinical outcomes.

The Flow Cytometry: Critical Assessment of Population Identification Methods (FlowCAP) challenges were established to compare the performance of computational methods for identifying cell populations in multidimensional flow cytometry data. Here we report the results of FlowCAP-IV where algorithms from seven different research groups predicted the time to progression to AIDS among a cohort of 384 HIV+ subjects, using antigen-stimulated peripheral blood mononuclear cell (PBMC) samples analyzed with a 14-color staining panel. Two approaches (FlowReMi.1 and flowDensity-flowType-RchyOptimyx) provided statistically significant predictive value in the blinded test set. Manual validation of submitted results indicated that unbiased analysis of single cell phenotypes could reveal unexpected cell types that correlated with outcomes of interest in high dimensional flow cytometry datasets.

FlowSOM: Using self-organizing maps for visualization and interpretation of cytometry data.

The number of markers measured in both flow and mass cytometry keeps increasing steadily. Although this provides a wealth of information, it becomes infeasible to analyze these datasets manually. When using 2D scatter plots, the number of possible plots increases exponentially with the number of markers and therefore, relevant information that is present in the data might be missed. In this article, we introduce a new visualization technique, called FlowSOM, which analyzes Flow or mass cytometry data using a Self-Organizing Map. Using a two-level clustering and star charts, our algorithm helps to obtain a clear overview of how all markers are behaving on all cells, and to detect subsets that might be missed otherwise. R code is available at https://github.com/SofieVG/FlowSOM and will be made available at Bioconductor.

Prediction of delayed graft function after kidney transplantation: comparison between logistic regression and machine learning methods.

BACKGROUND: Predictive models for delayed graft function (DGF) after kidney transplantation are usually developed using logistic regression. We want to evaluate the value of machine learning methods in the prediction of DGF. METHODS: 497 kidney transplantations from deceased donors at the Ghent University Hospital between 2005 and 2011 are included. A feature elimination procedure is applied to determine the optimal number of features, resulting in 20 selected parameters (24 parameters after conversion to indicator parameters) out of 55 retrospectively collected parameters. Subsequently, 9 distinct types of predictive models are fitted using the reduced data set: logistic regression (LR), linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), support vector machines (SVMs; using linear, radial basis function and polynomial kernels), decision tree (DT), random forest (RF), and stochastic gradient boosting (SGB). Performance of the models is assessed by computing sensitivity, positive predictive values and area under the receiver operating characteristic curve (AUROC) after 10-fold stratified cross-validation. AUROCs of the models are pairwise compared using Wilcoxon signed-rank test. RESULTS: The observed incidence of DGF is 12.5 %. DT is not able to discriminate between recipients with and without DGF (AUROC of 52.5 %) and is inferior to the other methods. SGB, RF and polynomial SVM are mainly able to identify recipients without DGF (AUROC of 77.2, 73.9 and 79.8 %, respectively) and only outperform DT. LDA, QDA, radial SVM and LR also have the ability to identify recipients with DGF, resulting in higher discriminative capacity (AUROC of 82.2, 79.6, 83.3 and 81.7 %, respectively), which outperforms DT and RF. Linear SVM has the highest discriminative capacity (AUROC of 84.3 %), outperforming each method, except for radial SVM, polynomial SVM and LDA. However, it is the only method superior to LR. CONCLUSIONS: The discriminative capacities of LDA, linear SVM, radial SVM and LR are the only ones above 80 %. None of the pairwise AUROC comparisons between these models is statistically significant, except linear SVM outperforming LR. Additionally, the sensitivity of linear SVM to identify recipients with DGF is amongst the three highest of all models. Due to both reasons, the authors believe that linear SVM is most appropriate to predict DGF.

[Genetic algorithm based multi-objective least square support vector machine for simultaneous determination of multiple components by near infrared spectroscopy].

The near infrared (NIR) spectrum contains a global signature of composition, and enables to predict different proper ties of the material. In the present paper, a genetic algorithm and an adaptive modeling technique were applied to build a multiobjective least square support vector machine (MLS-SVM), which was intended to simultaneously determine the concentrations of multiple components by NIR spectroscopy. Both the benchmark corn dataset and self-made Forsythia suspense dataset were used to test the proposed approach. Results show that a genetic algorithm combined with adaptive modeling allows to efficiently search the LS-SVM hyperparameter space. For the corn data, the performance of multi-objective LS-SVM was significantly better than models built with PLS1 and PLS2 algorithms. As for the Forsythia suspense data, the performance of multi-objective LS-SVM was equivalent to PLS1 and PLS2 models. In both datasets, the over-fitting phenomena were observed on RBFNN models. The single objective LS-SVM and MLS-SVM didn't show much difference, but the one-time modeling convenience al lows the potential application of MLS-SVM to multicomponent NIR analysis.

Machine-learning-based prediction of disability progression in multiple sclerosis: An observational, international, multi-center study.

BACKGROUND: Disability progression is a key milestone in the disease evolution of people with multiple sclerosis (PwMS). Prediction models of the probability of disability progression have not yet reached the level of trust needed to be adopted in the clinic. A common benchmark to assess model development in multiple sclerosis is also currently lacking. METHODS: Data of adult PwMS with a follow-up of at least three years from 146 MS centers, spread over 40 countries and collected by the MSBase consortium was used. With basic inclusion criteria for quality requirements, it represents a total of 15, 240 PwMS. External validation was performed and repeated five times to assess the significance of the results. Transparent Reporting for Individual Prognosis Or Diagnosis (TRIPOD) guidelines were followed. Confirmed disability progression after two years was predicted, with a confirmation window of six months. Only routinely collected variables were used such as the expanded disability status scale, treatment, relapse information, and MS course. To learn the probability of disability progression, state-of-the-art machine learning models were investigated. The discrimination performance of the models is evaluated with the area under the receiver operator curve (ROC-AUC) and under the precision recall curve (AUC-PR), and their calibration via the Brier score and the expected calibration error. All our preprocessing and model code are available at https://gitlab.com/edebrouwer/ms_benchmark, making this task an ideal benchmark for predicting disability progression in MS. FINDINGS: Machine learning models achieved a ROC-AUC of 0⋅71 ± 0⋅01, an AUC-PR of 0⋅26 ± 0⋅02, a Brier score of 0⋅1 ± 0⋅01 and an expected calibration error of 0⋅07 ± 0⋅04. The history of disability progression was identified as being more predictive for future disability progression than the treatment or relapses history. CONCLUSIONS: Good discrimination and calibration performance on an external validation set is achieved, using only routinely collected variables. This suggests machine-learning models can reliably inform clinicians about the future occurrence of progression and are mature for a clinical impact study.

Assessment of outdoor radiofrequency electromagnetic field exposure through hotspot localization using kriging-based sequential sampling.

In this study, a novel methodology is proposed to create heat maps that accurately pinpoint the outdoor locations with elevated exposure to radiofrequency electromagnetic fields (RF-EMF) in an extensive urban region (or, hotspots), and that would allow local authorities and epidemiologists to efficiently assess the locations and spectral composition of these hotspots, while at the same time developing a global picture of the exposure in the area. Moreover, no prior knowledge about the presence of radiofrequency radiation sources (e.g., base station parameters) is required. After building a surrogate model from the available data using kriging, the proposed method makes use of an iterative sampling strategy that selects new measurement locations at spots which are deemed to contain the most valuable information-inside hotspots or in search of them-based on the prediction uncertainty of the model. The method was tested and validated in an urban subarea of Ghent, Belgium with a size of approximately 1 km2. In total, 600 input and 50 validation measurements were performed using a broadband probe. Five hotspots were discovered and assessed, with maximum total electric-field strengths ranging from 1.3 to 3.1 V/m, satisfying the reference levels issued by the International Commission on Non-Ionizing Radiation Protection for exposure of the general public to RF-EMF. Spectrum analyzer measurements in these hotspots revealed five radiofrequency signals with a relevant contribution to the exposure. The radiofrequency radiation emitted by 900 MHz Global System for Mobile Communications (GSM) base stations was always dominant, with contributions ranging from 45% to 100%. Finally, validation of the subsequent surrogate models shows high prediction accuracy, with the final model featuring an average relative error of less than 2dB (factor 1.26 in electric-field strength), a correlation coefficient of 0.7, and a specificity of 0.96.

Exposure assessment of mobile phone base station radiation in an outdoor environment using sequential surrogate modeling.

Human exposure to background radiofrequency electromagnetic fields (RF-EMF) has been increasing with the introduction of new technologies. There is a definite need for the quantification of RF-EMF exposure but a robust exposure assessment is not yet possible, mainly due to the lack of a fast and efficient measurement procedure. In this article, a new procedure is proposed for accurately mapping the exposure to base station radiation in an outdoor environment based on surrogate modeling and sequential design, an entirely new approach in the domain of dosimetry for human RF exposure. We tested our procedure in an urban area of about 0.04 km(2) for Global System for Mobile Communications (GSM) technology at 900 MHz (GSM900) using a personal exposimeter. Fifty measurement locations were sufficient to obtain a coarse street exposure map, locating regions of high and low exposure; 70 measurement locations were sufficient to characterize the electric field distribution in the area and build an accurate predictive interpolation model. Hence, accurate GSM900 downlink outdoor exposure maps (for use in, e.g., governmental risk communication and epidemiological studies) are developed by combining the proven efficiency of sequential design with the speed of exposimeter measurements and their ease of handling.

Wideband parametric baseband macromodeling of linear and passive photonic circuits via complex vector fitting.

A novel wideband parametric baseband macromodeling technique for passive photonic devices and circuits is presented. It allows to efficiently estimate the baseband scattering representations of a linear, passive photonic system as a function of a set of design variables, such as geometrical layout or substrate features. The proposed technique relies on the interpolation of macromodels computed via a complex vector fitting (CVF) algorithm, by adopting a methodology based on amplitude and frequency scaling that preserves, by construction, the physical properties of the system, such as causality, stability and passivity. For a specified combination of the design parameters, a rational CVF model is derived that can be simulated by a wide range of ordinary differential equation (ODE) solvers or circuit simulators. Additionally, time-domain simulations using the computed model can be performed at arbitrary optical carrier frequencies, thus allowing for the simulation of multi-wavelength systems. Two application examples are presented to demonstrate the flexibility and advantages of the proposed method.

Split BiRNN for real-time activity recognition using radar and deep learning.

Radar systems can be used to perform human activity recognition in a privacy preserving manner. This can be achieved by using Deep Neural Networks, which are able to effectively process the complex radar data. Often these networks are large and do not scale well when processing a large amount of radar streams at once, for example when monitoring multiple rooms in a hospital. This work presents a framework that splits the processing of data in two parts. First, a forward Recurrent Neural Network (RNN) calculation is performed on an on-premise device (usually close to the radar sensor) which already gives a prediction of what activity is performed, and can be used for time-sensitive use-cases. Next, a part of the calculation and the prediction is sent to a more capable off-premise machine (most likely in the cloud or a data center) where a backward RNN calculation is performed that improves the previous prediction sent by the on-premise device. This enables fast notifications to staff if troublesome activities occur (such as falling) by the on-premise device, while the off-premise device captures activities missed or misclassified by the on-premise device.

An ontology-based nurse call management system (oNCS) with probabilistic priority assessment.

BACKGROUND: The current, place-oriented nurse call systems are very static. A patient can only make calls with a button which is fixed to a wall of a room. Moreover, the system does not take into account various factors specific to a situation. In the future, there will be an evolution to a mobile button for each patient so that they can walk around freely and still make calls. The system would become person-oriented and the available context information should be taken into account to assign the correct nurse to a call.The aim of this research is (1) the design of a software platform that supports the transition to mobile and wireless nurse call buttons in hospitals and residential care and (2) the design of a sophisticated nurse call algorithm. This algorithm dynamically adapts to the situation at hand by taking the profile information of staff members and patients into account. Additionally, the priority of a call probabilistically depends on the risk factors, assigned to a patient. METHODS: The ontology-based Nurse Call System (oNCS) was developed as an extension of a Context-Aware Service Platform. An ontology is used to manage the profile information. Rules implement the novel nurse call algorithm that takes all this information into account. Probabilistic reasoning algorithms are designed to determine the priority of a call based on the risk factors of the patient. RESULTS: The oNCS system is evaluated through a prototype implementation and simulations, based on a detailed dataset obtained from Ghent University Hospital. The arrival times of nurses at the location of a call, the workload distribution of calls amongst nurses and the assignment of priorities to calls are compared for the oNCS system and the current, place-oriented nurse call system. Additionally, the performance of the system is discussed. CONCLUSIONS: The execution time of the nurse call algorithm is on average 50.333 ms. Moreover, the oNCS system significantly improves the assignment of nurses to calls. Calls generally have a nurse present faster and the workload-distribution amongst the nurses improves.

Portable Detection of Apnea and Hypopnea Events Using Bio-Impedance of the Chest and Deep Learning.

Sleep apnea is one of the most common sleep-related breathing disorders. It is diagnosed through an overnight sleep study in a specialized sleep clinic. This setup is expensive and the number of beds and staff are limited, leading to a long waiting time. To enable more patients to be tested, and repeated monitoring for diagnosed patients, portable sleep monitoring devices are being developed. These devices automatically detect sleep apnea events in one or more respiration-related signals. There are multiple methods to measure respiration, with varying levels of signal quality and comfort for the patient. In this study, the potential of using the bio-impedance (bioZ) of the chest as a respiratory surrogate is analyzed. A novel portable device is presented, combined with a two-phase Long Short-Term Memory (LSTM) deep learning algorithm for automated event detection. The setup is benchmarked using simultaneous recordings of the device and the traditional polysomnography in 25 patients. The results demonstrate that using only the bioZ, an area under the precision-recall curve of 46.9% can be achieved, which is on par with automatic scoring using a polysomnography respiration channel. The sensitivity, specificity and accuracy are 58.4%, 76.2% and 72.8% respectively. This confirms the potential of using the bioZ device and deep learning algorithm for automatically detecting sleep respiration events during the night, in a portable and comfortable setup.

A voltage and current measurement dataset for plug load appliance identification in households.

This paper presents the Plug-Load Appliance Identification Dataset (PLAID), a labelled dataset containing records of the electrical voltage and current of domestic electrical appliances obtained at a high sampling frequency (30 kHz). The dataset contains 1876 records of individually-metered appliances from 17 different appliance types (e.g., refrigerators, microwave ovens, etc.) comprising 330 different makes and models, and collected at 65 different locations in Pittsburgh, Pennsylvania (USA). Additionally, PLAID contains 1314 records of the combined operation of 13 of these appliance types (i.e., measurements obtained when multiple appliances were active simultaneously). Identifying electrical appliances based on electrical measurements is of importance in demand-side management applications for the electrical power grid including automated load control, load scheduling and non-intrusive load monitoring. This paper provides a systematic description of the measurement setup and dataset so that it can be used to develop and benchmark new methods in these and other applications, and so that extensions to it can be developed and incorporated in a consistent manner.

Automated Sleep Apnea Detection in Raw Respiratory Signals Using Long Short-Term Memory Neural Networks.

Sleep apnea is one of the most common sleep disorders and the consequences of undiagnosed sleep apnea can be very severe, ranging from increased blood pressure to heart failure. However, many people are often unaware of their condition. The gold standard for diagnosing sleep apnea is an overnight polysomnography in a dedicated sleep laboratory. Yet, these tests are expensive and beds are limited as trained staff needs to analyze the entire recording. An automated detection method would allow a faster diagnosis and more patients to be analyzed. Most algorithms for automated sleep apnea detection use a set of human-engineered features, potentially missing important sleep apnea markers. In this paper, we present an algorithm based on state-of-the-art deep learning models for automatically extracting features and detecting sleep apnea events in respiratory signals. The algorithm is evaluated on the Sleep-Heart-Health-Study-1 dataset and provides per-epoch sensitivity and specificity scores comparable to the state of the art. Furthermore, when these predictions are mapped to the apnea-hypopnea index, a considerable improvement in per-patient scoring is achieved over conventional methods. This paper presents a powerful aid for trained staff to quickly diagnose sleep apnea.

Accurate prediction of blood culture outcome in the intensive care unit using long short-term memory neural networks.

INTRODUCTION: Blood cultures are often performed in the intensive care unit (ICU) to detect bloodstream infections and identify pathogen type, further guiding treatment. Early detection is essential, as a bloodstream infection can give cause to sepsis, a severe immune response associated with an increased risk of organ failure and death. PROBLEM STATEMENT: The early clinical detection of a bloodstream infection is challenging but rapid targeted treatment, within the first place antimicrobials, substantially increases survival chances. As blood cultures require time to incubate, early clinical detection using physiological signals combined with indicative lab values is pivotal. OBJECTIVE: In this work, a novel method is constructed and explored for the potential prediction of the outcome of a blood culture test. The approach is based on a temporal computational model which uses nine clinical parameters measured over time. METHODOLOGY: We use a bidirectional long short-term memory neural network, a type of recurrent neural network well suited for tasks where the time lag between a predictive event and outcome is unknown. Evaluation is performed using a novel high-quality database consisting of 2177 ICU admissions at the Ghent University Hospital located in Belgium. RESULTS: The network achieves, on average, an area under the receiver operating characteristic curve of 0.99 and an area under the precision-recall curve of 0.82. In addition, our results show that predicting several hours upfront is possible with only a small decrease in predictive power. In this setting, it outperforms traditional non-temporal, machine learning models. CONCLUSION: Our proposed computational model accurately predicts the outcome of blood culture tests using nine clinical parameters. Moreover, it can be used in the ICU as an early warning system to detect patients at risk of blood stream infection.

Systematic Comparison of Respiratory Signals for the Automated Detection of Sleep Apnea.

Sleep apnea is one of the most common sleep disorders. It is characterized by the cessation of breathing during sleep due to airway blockages (obstructive sleep apnea) or disturbances in the signals from the brain (central sleep apnea). The gold standard for diagnosing sleep apnea is performing an overnight polysomnography recording which contains, among others, a wide array of respiratory signals. Respiration information can also be extracted from other physiological signals such as an electrocardiogram or from a bio-impedance measurement on the chest. Studies have shown that algorithms can be developed for automated sleep apnea detection using one of these many respiratory signals. In this work, the predictive power of these different respiratory signals is analyzed and compared. The results provide useful insights into the comparative predictive power of the different respiratory signals in a realistic setting for automated sleep apnea detection and provide a basis for the development of less obtrusive measurement techniques.

Automated Assessment of Bone Age Using Deep Learning and Gaussian Process Regression.

Bone age is an essential measure of skeletal maturity in children with growth disorders. It is typically assessed by a trained physician using radiographs of the hand and a reference model. However, it has been described that the reference models leave room for interpretation leading to a large inter-observer and intra-observer variation. In this work, we explore a novel method for automated bone age assessment to assist physicians with their estimation. It consists of a powerful combination of deep learning and Gaussian process regression. Using this combination, sensitivity of the deep learning model to rotations and flips of the input images can be exploited to increase overall predictive performance compared to only using the deep learning network. We validate our approach retrospectively on a set of 12611 radiographs of patients between 0 and 19 years of age.

Adaptive initial step size selection for Simultaneous Perturbation Stochastic Approximation.

A difficulty in using Simultaneous Perturbation Stochastics Approximation (SPSA) is its performance sensitivity to the step sizes chosen at the initial stage of the iteration. If the step size is too large, the solution estimate may fail to converge. The proposed adaptive stepping method automatically reduces the initial step size of the SPSA so that reduction of the objective function value occurs more reliably. Ten mathematical functions each with three different noise levels were used to empirically show the effectiveness of the proposed idea. A parameter estimation example of a nonlinear dynamical system is also included.