Heart Risk Prediction System Based on Supervised ANN

Hypertensive disorders of pregnancy (HDP) cause adverse maternal and fetal outcomes, and establishing an early prediction method for HDP is needed. Existing methods based on serum markers require blood sampling. Therefore, we aimed to develop and validate a noninvasive HDP prediction model based on home blood pressure monitoring. In a development cohort, home blood pressure monitoring data from 443 pregnant women including 65 HDPs were divided into training data (n=365) and test data (n=78) to develop a logistic regression-based prediction model. Normal blood pressure variations depending on season and gestational age were subtracted. At each time point, 4 features were calculated for the last 4 weeks of data: average systolic blood pressure, average diastolic blood pressure, correlation coefficient between systolic blood pressure and diastolic blood pressure, and upward trend of systolic blood pressure against day. In a validation cohort, HBPM data from 264 pregnant women including 33 HDPs were collected prospectively and used to validate the model. The area under the receiver operating characteristic curve was 0.949 (95% CI, 0.890-1.000), 0.884 (95% CI, 0.807-0.961), and 0.845 (95% CI, 0.759-0.930) for the training, test, and validation data, respectively. Sensitivity, specificity, positive predictive value, and negative predictive value were 0.923, 0.888, 0.387, and 0.993 for the training data; 0.667, 0.897, 0.867, and 0.729 for the test data; 0.758, 0.766, 0.316, and 0.957 for the validation data. Our HDP prediction model based on home blood pressure monitoring showed high negative predictive values and may contribute to reducing medical consultations.

Passive sensing applications are limited by their inability to determine who is using a device, a critical concern in child mobile device use research, where devices are often shared between siblings or between a child and their parent. Our previous work leveraged behavioral biometrics to identify a target child user; however, it is unknown what type of training data is necessary for optimal model performance. This study evaluated model performance across different characteristics of training data. Thirty-six children (11.3 ± 0.9 years, 56% female) self-selected a video or a game on iPads for 10 min while laying and for another 5 min while sitting. The SensorLog application captured iPad accelerometer and gyroscope data while the child interacted with the device. Machine learning algorithms including Neural Network (NN), Random Forest (RF), k-Nearest Neighbors (k-NN), and SwipeFormer were applied to determine the most important aspects of training data to optimize model performance. The aspects of training data evaluated included (1) varying the length (i.e., seconds of training data), (2) varying the user position (i.e., sitting, laying), and (3) varying the time proximity between training and testing data. F1 score was used to evaluate model performance. The SwipeFormer F1 scores were lowest when the training data was further from the test data (0 when training data was 11 min away from test data) and highest when training data was close to test data (0.91 when training data was the minute preceding test data). The SwipeFormer F1 scores were highest when predicting the user laying from laying (0.97) and sitting from sitting (0.94), and lowest when predicting the user sitting from laying (0) and laying from sitting (0). The length of training data had little impact on performance, with a SwipeFormer F1 score of 0.91 when training on one minute of data and a SwipeFormer F1 score of 0.94 when training on twelve minutes of data. Because researchers would likely be predicting users at different timepoints than their training data, research should focus on improving model performance for identifying users independent of time proximity for training and test data.

Spatial application of Random Forest models for fine-scale coastal vegetation classification using object based analysis of aerial orthophoto and DEM data

The purpose of this study is to predict the decade rainfall in Semarang City using two main methods, namely Seasonal Autoregressive Integrated Moving Average with Exogenous Variables (SARIMAX) and Long Short Term Memory (LSTM). The methodology of this study begins with data pre-processing, which includes data deletion analysis using dropna and data normalization using Min-Max Scaling to reduce scale the to between 0 and 1. The dataset is then divided into 80% training data and 20% test data. The validity of the data (X_test, Y_test) using the best 56-epoch data validation (val_loss) is better than the validity of the training data (loss). On the other hand, SARIMAX uses the (2,1,2), (2,1,2,36) model, and its validation techniques include Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Coefficient of Determination (R2). Specifically, the RMSE of the LSTM model is 19.6, and its RMSE of the SARIMAX is 31.05. The MAE of LSTM is 15.0, SARIMAX is 24.5, and the R2 of LSTM is 0.814, and SARIMAX is 0.52. Lower RMSE and MAE values indicate lower prediction errors, but a higher R2 value of 1 indicates that LSTM can explain 81% of the actual data variation, which is better than SARIMAX, which is only about 52%. The main finding of this study is that the LSTM model performs better when recommending rainfall datasets.

Financial institutions have a large amount of data on their borrowers, which can be used to predict the probability of borrowers defaulting their loan or not. Some of the models that have been used to predict individual loan defaults include linear discriminant analysis models and extreme value theory models. These models are parametric in nature since they assume that the response being investigated takes a particular functional form. However, there is a possibility that the functional form used to estimate the response is very different from the actual functional form of the response. The purpose of this research was to analyze individual loan defaults in Kenya using the logistic regression model. The data used in this study was obtained from equity bank of Kenya for the period between 2006 to 2016. A random sample of 1000 loan applicants whose loans had been approved by equity bank of Kenya during this period was obtained. Data obtained was on the credit history, purpose of the loan, loan amount, nature of the saving account, employment status, sex of the applicant, age of the applicant, security used when acquiring the loan and the area of residence of the applicant (rural or urban). This study employed a quantitative research design, it deals with individual loans defaults as group characteristics of a borrower. The data was pre-processed by seeding using R- Software and then split into training dataset and test data set. The train data was used to train the logistic regression model by employing Supervised machine learning approach. The R-statistical software was used for the analysis of the data. The test data set was used to do cross-validation of the developed logistic model which later was used for analysis prediction of individual loan defaults. This study focused on the analysis of individual loan defaults in Kenya using the logistic regression model in Machine learning. The logistic regression model predicted 303 defaults from train data set, 122 non-defaults and misclassified loans were 56 and 69. The model had an accuracy of 0.7727 with the train data and 0.7333 with the test data. The logistic regression model showed a precision of 0.8440 and 0.8244 with the train and test data respectively. The performance of the model with both the train and test data was illustrated using a plot of train errors and test errors against sample size on the same axes. The plot showed that the performance of the model increases with an increase in sample size. The study recommended the use of logistic regression in conjunction with supervised machine learning approach in loan default prediction in financial institutions and also more research should be carried out on ensemble methods of loan defaults prediction in order to increase the prediction accuracy.

Background: Despite the increasing understanding of the molecular characteristics of cancer, chemotherapy success rates remain low for many cancer types. Studies have attempted to identify patient and tumor characteristics that predict sensitivity or resistance to different types of conventional chemotherapies, yet a concise model that predicts chemosensitivity based on gene expression signatures across cancer types remains to be formulated. We attempted to generate a pan-cancer chemosensitivity predictive model using publicly available data from multiple sources. Such a model may increase the likelihood of identifying the type of chemotherapy most likely to succeed for a given patient based on the gene expression signature of their tumor. Methods: Data used to build the predictive model were obtained from the Genomics of Drug Sensitivity in Cancer (GDSC) database, consisting of gene expression profiles from 962 cancer cell lines via RNA sequencing (RNA-seq) and drug sensitivity profiles reported as ln(IC50). Predictive gene signatures were generated using a cross-validated generalized linear model (leave one out cross-validation) using elasticnet penalization parameters. Models were generated for each individual drug tested by the GDSC cohort, as well as different classes of chemotherapeutics (platinum agents, topoisomerase inhibitors, mustard agents, antibiotics and anti-fungals, anti-metabolites, and taxanes). Accuracy of the models was determined using normalized mean square error (nRMSE). Models were then validated using publicly available data from Cancer Cell Line Encyclopedia (CCLE), NCI-60, and the Patient-Derived Xenograft (PDX) Clinical Trial (PCT) database. Models were further validated using human tumor datasets available via the Gene Expression Omnibus (GEO). As the training data used to generate the models were from RNA-seq, and some of the testing and validation data were generated using microarray technology, feature-specific quantile normalization was used to enable cross-platform analyses. Results: For most single-drug gene signatures, accuracy measured by nRMSE ranged from 0.10-0.20, which suggests that for any given model the root mean squared error is 10-20% of the range of actual ln(IC50) in the tested data. Chemotherapy class-level models yielded slightly less accuracy, with nRMSE ranging from 0.15-0.25 for most classes. When considering how well the models predicted chemosensitivity within cancer types, accuracy was improved in some cancer types (e.g., lung cancer and head and neck cancer), with more heterogeneous cancer types (e.g., breast cancer) giving less accuracy. Conclusions: Our results show that the models generated can predict chemosensitivity across cancer types with clinical useful levels of accuracy, with some cancer types resulting in a high rate of accuracy across several classes of chemotherapy. The inclusion of future datasets, particularly from those cancer types in which chemosensitivity has been difficult to predict, may provide opportunities to strengthen model accuracy as well as decrease the numbers of genes needed to assess chemosensitivity. Citation Format: Wells JD, Miller TW. Development of pan-cancer transcriptional signatures that predict chemosensitivity [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P3-11-15.

Nitrate is a common pollutant in surface waters which water companies must monitor for regulatory and safety reasons. The presence of nitrate in deionised water is detected and concentration estimated from microwave spectroscopy measurements in the range 9kHz-6GHz. Experimental results were obtained for 19 solutions (18 salt solutions in deionised water and 1 deionised water), each measured 10 times with 4001 points (total N=190). The resulting data was randomly assigned into equal parts training and test data (N=95 each). Both regression (for the estimation of nitrate concentration) and classification (for detecting the presence of nitrate) methods were considered, with a rigorous feature selection procedure used to identify two frequencies for each of the classification and regression problems. For detection classification models were applied with nitrate levels binned using 30mg/l as the threshold. A logistic regression model achieved AUROC of 0.9875 on test data and a multi-layer perceptron achieved AUROC of 0.9871. In each case the positive predictive value of the model could be optimised at 100% with sensitivity of 90% and specificity of 100%. For the concentration estimates, a linear regression model was able to explain 42% of the variance in the training data and 45% of the variance in the test data and an MLP model delivered similar performance, explaining 43% of variance in the training data and 47% of variance in the test data. A sensor based on this model would be appropriate for detecting the presence of nitrate above a given threshold but poor at estimating concentration.

This focus issue of Physiological Measurement follows the 38th Annual International Computing in Cardiology (CinC) Conference, hosted in Hangzhou, China in September 2011 by Zhejiang University. Each year, the NIH-sponsored PhysioNet resource (http://physionet.org/) runs an open competition lasting several months, aimed at encouraging the development of solutions to an unsolved or poorly solved problem in biomedicine, in most cases making use of relevant clinical and experimental data provided freely by PhysioNet. Participants in these annual challenges discuss their diverse approaches to the Challenge problems during dedicated scientific sessions at CinC. The topics of these PhysioNet/CinC Challenges range from physiologic signal processing and analysis to forecasting and modelling clinically important events and processes.

Extreme weather modeling is a challenge for modeling experts in Indonesia and the world. Extreme weather prediction is a complex problem because the chances of it happening are very small, so the developed models often have a low level of accuracy. The purpose of this research is to combine the classic model, Autoregressive Integrated Moving Average (ARIMA), recurrent neural network (RNN) model using Adam and SGD estimation (RNN-Adam and RNN-SGD) with the reLU, tanh, sigmoid and gaussian activation functions. In addition, the ARIMA-RNN mix model was also demonstrated in this study. These models are applied to monthly period extreme weather data obtained from the Meteorology, Climatology and Geophysics Agency (BMKG) of West Sulawesi Province which are converted into training data and test data. The RMSE value is used to see the level of prediction accuracy in both training data and test data. Based on the research results, the best model obtained for modeling Indonesia’s extreme weather is the ARIMA-RNN-Adam mix model with the reLU activation function based on the RMSE value on the training and test data. At n = 50, the smallest RMSE and MSE values of the third model are the ARIMA-RNN-Adam model which is 0.23212 using the reLU activation function, then the ARIMA-RNN-SGD model which is 0.25432 with the same activation function, while the ARIMA value is 0.3270. At n=100 it can be seen that the smallest RMSE and MSE values of the three models are the ARIMA-RNN-Adam model which is equal to 0.25149 using the reLU activation function, then the ARIMA-RNN-SGD model which is equal to 0.25256 with the same activation function, while the ARIMA value is 0.2644.

Automated monitoring of the movements and behaviour of animals is a valuable research tool. Recently, machine learning tools were applied to many species to classify units of behaviour. For the monitoring of wild species, collecting enough data for training models might be problematic, thus we examine how machine learning models trained on one species can be applied to another closely related species with similar behavioural conformation. We contrast two ways to calculate accuracies, termed here as overall and threshold accuracy, because the field has yet to define solid standards for reporting and measuring classification performances. We measure 21 dogs and 7 wolves, and find that overall accuracies are between 51 and 60% for classifying 8 behaviours (lay, sit, stand, walk, trot, run, eat, drink) when training and testing data are from the same species and between 41 and 51% when training and testing is cross-species. We show that using data from dogs to predict the behaviour of wolves is feasible. We also show that optimising the model for overall accuracy leads to similar overall and threshold accuracies, while optimizing for threshold accuracy leads to threshold accuracies well above 80%, but yielding very low overall accuracies, often below the chance level. Moreover, we show that the most common method for dividing the data between training and testing data (random selection of test data) overestimates the accuracy of models when applied to data of new specimens. Consequently, we argue that for the most common goals of animal behaviour recognition overall accuracy should be the preferred metric. Considering, that often the goal is to collect movement data without other methods of observation, we argue that training data and testing data should be divided by individual and not randomly.

— The current Governor of DKI Jakarta, even though he has been elected since 2017 is always interesting to talk about or even comment on. Comments that appear come from the media directly or through social media. Twitter has become one of the social media that is often used as a media to comment on elected governors and can even become a trending topic on Twitter social media. Netizens who comment are also varied, some are always Tweeting criticism, some are commenting Positively, and some are only re-Tweeting. In this research, a prediction of whether active Netizens will tend to always lead to Positive or Negative comments will be carried out in this study. Model algorithms used are Decision Tree, Naïve Bayes, Random Forest, and also Ensemble. Twitter data that is processed must go through preprocessing first before proceeding using Rapidminer. In trials using Rapidminer conducted in four trials by dividing into two parts, namely testing data and training data. Comparisons made are 10% testing data: 90% Training data, then 20% testing data: 80% training data, then 30% testing data: 70% training data, and the last is 35% testing data: 65% training data. The average Accuracy for the Decision Tree algorithm is 93.15%, while for the Naïve Bayes algorithm the Accuracy is 91.55%, then for the Random Forest algorithm is 93.41, and the last is the Ensemble algorithm with an Accuracy of 93, 42%. here. Keywords — Decision Tree, Naïve Bayes, Random Forest, Set, Twitter.

Statistical learning algorithms strongly rely on an oversimplified assumption for optimal performance, that is, source (training) and target (testing) data are independent and identically distributed. Variation in human tissue, physician labeling and physical imaging parameters (PIPs) in the generative process, yield medical image datasets with statistics that render this central assumption false. When deploying models, new examples are often out of distribution with respect to training data, thus, training robust dependable and predictive models is still a challenge in medical imaging with significant accuracy drops common for deployed models. This statistical variation between training and testing data is referred to as domain shift (DS).To the best of our knowledge we provide the first empirical evidence that variation in PIPs between test and train medical image datasets is a significant driver of DS and model generalization error is correlated with this variance. We show significant covariate shift occurs due to a selection bias in sampling from a small area of PIP space for both inter and intra-hospital regimes. In order to show this, we control for population shift, prevalence shift, data selection biases and annotation biases to investigate the sole effect of the physical generation process on model generalization for a proxy task of age group estimation on a combined 44k image mammogram dataset collected from five hospitals.We hypothesize that training data should be sampled evenly from PIP space to produce the most robust models and hope this study provides motivation to retain medical image generation metadata that is almost always discarded or redacted in open source datasets. This metadata measured with standard international units can provide a universal regularizing anchor between distributions generated across the world for all current and future imaging modalities.

Website Fingerprinting (WF) attacks pose a serious threat to users' online privacy, including for users of the Tor anonymity system. By exploiting recent advances in deep learning, WF attacks like Deep Fingerprinting (DF) have reached up to 98% accuracy. The DF attack, however, requires large amounts of training data that needs to be updated regularly, making it less practical for the weaker attacker model typically assumed in WF. Moreover, research on WF attacks has been criticized for not demonstrating attack effectiveness under more realistic and more challenging scenarios. Most research on WF attacks assumes that the testing and training data have similar distributions and are collected from the same type of network at about the same time. In this paper, we examine how an attacker could leverage N-shot learning---a machine learning technique requiring just a few training samples to identify a given class---to reduce the effort of gathering and training with a large WF dataset as well as mitigate the adverse effects of dealing with different network conditions. In particular, we propose a new WF attack called Triplet Fingerprinting (TF) that uses triplet networks for N-shot learning. We evaluate this attack in challenging settings such as where the training and testing data are collected multiple years apart on different networks, and we find that the TF attack remains effective in such settings with 85% accuracy or better. We also show that the TF attack is also effective in the open world and outperforms traditional transfer learning. On top of that, the attack requires only five examples to recognize a website, making it dangerous in a wide variety of scenarios where gathering and training on a complete dataset would be impractical.

SummaryAccurate prediction of the solubility of chemical substances in solvents remains a challenge. The sparsity of high-quality solubility data is recognized as the biggest hurdle in the development of robust data-driven methods for practical use. Nonetheless, the effects of the quality and quantity of data on aqueous solubility predictions have not yet been scrutinized. In this study, the roles of the size and the quality of data sets on the performances of the solubility prediction models are unraveled, and the concepts of actual and observed performances are introduced. In an effort to curtail the gap between actual and observed performances, a quality-oriented data selection method, which evaluates the quality of data and extracts the most accurate part of it through statistical validation, is designed. Applying this method on the largest publicly available solubility database and using a consensus machine learning approach, a top-performing solubility prediction model is achieved.

MOTIVATION Glioblastoma is the most common and aggressive adult brain tumor. Clinical histopathologic evaluation is essential for tumor classification, which according to the World Health Organization is associated with prognostic information. Accurate prediction of patient overall survival (OS) from clinical routine baseline histopathology whole slide images (WSI) using advanced computational methods, while considering variations in the staining process, could contribute to clinical decision-making and patient management optimization. METHODS We utilize The Cancer Genome Atlas glioblastoma (TCGA-GBM) collection, comprising multi-institutional hematoxylin and eosin (H&E) stained frozen top-section WSI, genomic, and clinical data from 121 subjects. Data are randomly split into training (80%), validation (10%), and testing (10%) sets, while proportionally keeping the ratio of censored patients. We propose a novel deep learning algorithm to identify survival-discriminative histopathological patterns in a WSI, through feature maps, and quantitatively integrate them with gene expression and clinical data to predict patient OS. The concordance index (C-index) is used to quantify the predictive OS performance. Variations in slide staining are assessed through a novel population-based stain normalization approach, informed of glioblastoma distinct histologic sub-regions and their appearance from 509 H&E stained slides with corresponding anatomical annotations from the Ivy Glioblastoma Atlas Project (IvyGAP). RESULTS C-index was equal to 0.797, 0.713, and 0.703 for the training, validation, and testing data, respectively, prior to stain normalization. Following normalization, staining variations in H&E and ‘E’ gained significant improvements in IvyGAP (pWilcoxon< 0.01) and TCGA-GBM (pWilcoxon< 0.0001) data, respectively. These improvements contributed to further optimizing the C-index to 0.871, 0.777, and 0.780 for the training, validation, and testing data, respectively. CONCLUSIONS Appropriate normalization and integrative deep learning yield accurate OS prediction of glioblastoma patients through H&E slides, generalizable in multi-institutional data, potentially contributing to patient stratification in clinical trials. Our computationally-identified survival-discriminative histopathological patterns can contribute in further understanding glioblastoma.