scispace - formally typeset
Search or ask a question
Book ChapterDOI

Predicting the Risk of Readmission of Diabetic Patients Using Deep Neural Networks

01 Jan 2019-pp 385-392
TL;DR: This paper predicts whether a patient discharged from the hospital will return within 30 days or not, and builds a deep neural network based on a specific and optimized sequential architecture.
Abstract: One of the major concerns these days are hospital readmissions, as they are expensive and mark the shortfalls in health care. The United States alone has spent 41.3 billion dollars between January and November 2011 to treat patients readmitted within 30 days of discharge, according to the Agency for Healthcare Research and Quality (AHRQ). There are already several attempts made in the same field using Machine learning methods to leverage public health data to build a system for identifying diabetic patients facing a high risk of future readmission. This paper predicts whether a patient discharged from the hospital will return within 30 days or not. The best possible feature engineering pipeline is chosen to process the data so that it can be learnt by the model in the best manner in determining the most important evaluation metric. A classifier is built using the traditional machine learning algorithms such as Linear SVM, Random Forest. We have also built a deep neural network based on a specific and optimized sequential architecture.
Citations
More filters
Journal ArticleDOI
TL;DR: In this paper, the authors used Machine Leaning algorithms to predict if a patient who suffers from diabetes will be readmitted, after being discharged, using data about patients with diabetes, which is a chronic condition that occurs when the body does not produce enough or any insulin.
Abstract: Hospitals generate large amounts of data on a daily basis, but most of the time that data is just an overwhelming amount of information which never transitions to knowledge. Through the application of Data Mining techniques it is possible to find hidden relations or patterns among the data and convert those into knowledge that can further be used to aid in the decision-making of hospital professionals. This study aims to use information about patients with diabetes, which is a chronic (long-term) condition that occurs when the body does not produce enough or any insulin. The main purpose is to help hospitals improve their care with diabetic patients and consequently reduce readmission costs. An hospital readmission is an episode in which a patient discharged from a hospital is admitted again within a specified period of time (usually a 30 day period). This period allows hospitals to verify that their services are being performed correctly and also to verify the costs of these re-admissions. The goal of the study is to predict if a patient who suffers from diabetes will be readmitted, after being discharged, using Machine Leaning algorithms. The final results revealed that the most efficient algorithm was Random Forest with 0.898 of accuracy.

9 citations

Journal ArticleDOI
TL;DR: In this article, the authors proposed a group model to predict readmission by choosing between machine learning and deep learning algorithms based on performance improvement, which aims to achieve the best rate of between prediction rate accuracy for hospital readmission at the same time minimizing resources such as time delay and energy consumption.
Abstract: Readmission to the hospital is an important and critical procedure for the quality of health care as it is very costly and helps in determining the quality level of the point of care provided by the hospital to the patient. This paper proposes a group model to predict readmission by choosing between Machine Learning and Deep Learning algorithms based on performance improvement. The algorithms used for Machine Learning are Logistic Regression, K-Nearest Neighbors, and Support Vector Machine, while the algorithms used for Deep Learning are a Convolutional Neural Network and Recurrent Neural Network. The reasons for the appearance of the efficiency of the model depend on the are preparation of correct parameters and the values that control the learning. This paper aims to enhance the performance of both machine learning and deep learning based readmission models using hyperparameter optimization in both Personal Computer environments and Mobile Cloud Computing systems. The proposed model is called improving detection diabetic using hyperparameter optimization , the proposed model aims to achieve the best rate of between prediction rate accuracy for hospital readmission at the same time minimizing resources such as time delay and energy consumption. Results achieved by proposed model for Logistic Regression, K-Nearest Neighbors, and Support Vector Machine are (accuracy=0.671, 0.883, 0.901, time delay=5, 7, 20, and energy consumed=25, 32, 48) respectively, for Recurrent Neural Network and Convolutional Neural Network are (accuracy=0.854, 0.963, time delay=25, 660 energy consumed=89, 895) respectively. However, this proposed model takes a lot of time and energy consumed especially in Convolutional Neural Network. So, the experiments were conducted again, but in the cloud environment, based on the existence of two types of storage to preserve the accuracy but decreasing time and energy, the proposed model in cloud environment achieve for Logistic Regression, K-Nearest Neighbors, and Support Vector Machine (accuracy=0.671, 0.883, 0.901, time delay=2, 3, 8, and energy consumed=8, 9, 11) respectively, for Recurrent Neural Network, Convolutional Neural Network (accuracy=0.854, 0.963, time delay=15, 220, and energy consumed=20, 301) respectively.
References
More filters
Journal ArticleDOI
01 Oct 2001
TL;DR: Internal estimates monitor error, strength, and correlation and these are used to show the response to increasing the number of features used in the forest, and are also applicable to regression.
Abstract: Random forests are a combination of tree predictors such that each tree depends on the values of a random vector sampled independently and with the same distribution for all trees in the forest. The generalization error for forests converges a.s. to a limit as the number of trees in the forest becomes large. The generalization error of a forest of tree classifiers depends on the strength of the individual trees in the forest and the correlation between them. Using a random selection of features to split each node yields error rates that compare favorably to Adaboost (Y. Freund & R. Schapire, Machine Learning: Proceedings of the Thirteenth International conference, aaa, 148–156), but are more robust with respect to noise. Internal estimates monitor error, strength, and correlation and these are used to show the response to increasing the number of features used in the splitting. Internal estimates are also used to measure variable importance. These ideas are also applicable to regression.

79,257 citations

Proceedings Article
03 Dec 2012
TL;DR: The state-of-the-art performance of CNNs was achieved by Deep Convolutional Neural Networks (DCNNs) as discussed by the authors, which consists of five convolutional layers, some of which are followed by max-pooling layers, and three fully-connected layers with a final 1000-way softmax.
Abstract: We trained a large, deep convolutional neural network to classify the 1.2 million high-resolution images in the ImageNet LSVRC-2010 contest into the 1000 different classes. On the test data, we achieved top-1 and top-5 error rates of 37.5% and 17.0% which is considerably better than the previous state-of-the-art. The neural network, which has 60 million parameters and 650,000 neurons, consists of five convolutional layers, some of which are followed by max-pooling layers, and three fully-connected layers with a final 1000-way softmax. To make training faster, we used non-saturating neurons and a very efficient GPU implementation of the convolution operation. To reduce overriding in the fully-connected layers we employed a recently-developed regularization method called "dropout" that proved to be very effective. We also entered a variant of this model in the ILSVRC-2012 competition and achieved a winning top-5 test error rate of 15.3%, compared to 26.2% achieved by the second-best entry.

73,978 citations

Journal ArticleDOI
TL;DR: This paper provides an introduction to the WEKA workbench, reviews the history of the project, and, in light of the recent 3.6 stable release, briefly discusses what has been added since the last stable version (Weka 3.4) released in 2003.
Abstract: More than twelve years have elapsed since the first public release of WEKA. In that time, the software has been rewritten entirely from scratch, evolved substantially and now accompanies a text on data mining [35]. These days, WEKA enjoys widespread acceptance in both academia and business, has an active community, and has been downloaded more than 1.4 million times since being placed on Source-Forge in April 2000. This paper provides an introduction to the WEKA workbench, reviews the history of the project, and, in light of the recent 3.6 stable release, briefly discusses what has been added since the last stable version (Weka 3.4) released in 2003.

19,603 citations

01 Jan 2007

17,341 citations

Journal ArticleDOI
01 Aug 1997
TL;DR: The model studied can be interpreted as a broad, abstract extension of the well-studied on-line prediction model to a general decision-theoretic setting, and it is shown that the multiplicative weight-update Littlestone?Warmuth rule can be adapted to this model, yielding bounds that are slightly weaker in some cases, but applicable to a considerably more general class of learning problems.
Abstract: In the first part of the paper we consider the problem of dynamically apportioning resources among a set of options in a worst-case on-line framework. The model we study can be interpreted as a broad, abstract extension of the well-studied on-line prediction model to a general decision-theoretic setting. We show that the multiplicative weight-update Littlestone?Warmuth rule can be adapted to this model, yielding bounds that are slightly weaker in some cases, but applicable to a considerably more general class of learning problems. We show how the resulting learning algorithm can be applied to a variety of problems, including gambling, multiple-outcome prediction, repeated games, and prediction of points in Rn. In the second part of the paper we apply the multiplicative weight-update technique to derive a new boosting algorithm. This boosting algorithm does not require any prior knowledge about the performance of the weak learning algorithm. We also study generalizations of the new boosting algorithm to the problem of learning functions whose range, rather than being binary, is an arbitrary finite set or a bounded segment of the real line.

15,813 citations