Ensemble Of Feed-forward Neural Networks Research Articles

A fusion of data science and feed-forward neural network-based modelling of COVID-19 outbreak forecasting in IRAQ

BackgroundIraq is among the countries affected by the COVID-19 pandemic. As of 2 August 2020, 129,151 COVID-19 cases were confirmed, including 91,949 recovered cases and 4,867 deaths. After the announcement of lockdown in early April 2020, situation in Iraq was getting steady until late May 2020, when daily COVID-19 infections have raised suddenly due to gradual easing of lockdown restrictions. In this context, it is important to develop a forecasting model to evaluate the COVID-19 outbreak in Iraq and so to guide future health policy. MethodsCOVID-19 lag data were made available by the University of Anbar through their online analytical platform (https://www.uoanbar.edu.iq/covid/), engaged with the day-to-day figures form the Iraqi health authorities. 154 days of patient data were provided covering the period from 2 March 2020 to 2 August 2020. An ensemble of feed-forward neural networks has been adopted to forecast COVID-19 outbreak in Iraq. Also, this study highlights some key questions about this pandemic using data analytics. ResultsForecasting were achieved with accuracy of 87.6% for daily infections, 82.4% for daily recovered cases, and 84.3% for daily deaths. It is anticipated that COVID-19 infections in Iraq will reach about 308,996 cases by the end of September 2020, including 228,551 to recover and 9,477 deaths. ConclusionThe applications of artificial neural networks supported by advanced data analytics represent a promising solution through which to realise intelligent solutions, enabling the space of analytical operations to drive a national health policy to contain COVID-19 pandemic.

Validation of the Genetically-Defined DLBCL Subtypes and Generation of a Parsimonious Probabilistic Classifier

Diffuse large B-cell lymphoma (DLBCL) is a clinically and molecularly heterogeneous disease with recognized transcriptional subtypes associated with normal cells of origin, activated B-cell (ABC) and germinal center B-cell (GCB) tumors. Emerging data suggested that additional heterogeneity existed, prompting us to comprehensively characterize genomic signatures of 304 newly diagnosed DLBCLs from patients treated with state-of-the-art therapy. We integrated recurrent mutations, somatic copy number alterations (SCNAs) and structural variants (SVs) and identified 5 genetically distinct DLBCL clusters (C1- C5 DLBCLs; Chapuy, Stewart, Dunford, et al. Nat Med 2018). Specifically, we identified two genetically distinct ABC subtypes, including favorable-risk C1 DLBCLs with features of extrafollicular origin and alterations also seen in transformed marginal zone lymphomas (NOTCH2 and NF-κB pathway member mutations and BCL6 SVs). Unfavorable-risk C5 ABC DLBCLs harbored frequent 18q/BCL2 copy gain and co-occurring CD79B and MYD88L265P mutations. We also identified two genetically distinct GCB subtypes, including unfavorable-risk C3 DLBCLs with frequent BCL2 SVs, mutations in chromatin-modifying enzymes (CREBBP, MLL2, EZH2) and BCR/PI3K signaling pathway members (including inactivating PTEN mutations and copy loss). Favorable-risk C4 GCB DLBCLs had frequent mutations in core and linker histones and signaling intermediates (SGK1, BRAF and STAT3). Additionally, we identified an ABC/GCB-independent subtype, C2 DLBCLs, characterized by frequent bi-allelic TP53 inactivation, 9p21.23/CDKN2A copy loss and associated genomic instability reflected in recurrent SCNAs, increased genome doublings and a distinct outcome following induction therapy. A next step in utilizing the characterized genetic substructure was to confirm it in an independent series and develop a molecular classifier that allows prospective identification of C1-C5 DLBCLs. To this end, we accessed whole exome sequencing, copy number and SV data from a recent cohort of newly diagnosed DLBCLs (39 tumor-normal pairs, 462 tumor-only samples; Schmitz et al. NEJM 2018). All samples were re-analyzed using our mutational and SCNA pipelines and our newly generated tumor-only algorithm (Chapuy, Stewart, Dunford, et al. Nat Med 2018) to avoid batch effects and harmonize the datasets. SVs were used as reported. Purity and ploidy were inferred using ABSOLUTE and samples with missing data or low purity were removed. For the combined cohort (579 samples), we assessed our previously characterized 158 genetic drivers (Chapuy, Stewart, Dunford, et al. Nat Med 2018) and confirmed equal distribution of their marginal frequencies (R=0.88, p=1.5e-51), excluding batch effects. Next, we applied non-negative matrix factorization (NNF) consensus clustering to the combined dataset (158 genetic drivers vs. 579 tumors) and confirmed the C1-C5 DLBCL genetic clusters. Notably, tumors from both series contributed at comparable frequencies to the respective C1-C5 DLBCLs. We also noted an enrichment of the alternative genetic labels from Schmitz et al. in 3 of our C1-C5 DLBCL subtypes (B2N in C1 DLBCLs, p<0.0001; EZB in C3 DLBCLs, p<0.0001; MCD in C5 DLBCLs, p<0.0001). These data confirmed the identity of the C1-C5 DLBCL clusters in an independent cohort. Next, we developed a molecular classifier that prospectively identified C1-C5 DLBCLs using a minimum number of easy-to-measure features. The NMF-defined classes of the combined cohort were used as gold-standard training and validation datasets. We tested different models for classification and selected an artificial neural network approach which provides accurate classification of individual samples and well-calibrated confidence metrics. To minimize potential overtraining, we developed a reduced input feature set of the 22 most discriminating features, constructed confidence metrics for each sample and trained an ensemble of Feed-Forward Neural Networks via 10-fold cross validation. With this approach, our classifier had 84% accuracy for the total set and 94% accuracy for the high-confidence samples (70% of all samples). The newly developed parsimonious classifier will allow prospective identification of the independently confirmed C1-C5 DLBCL subtypes in newly diagnosed patients, a necessity for clinical application. Disclosures Getz: Pharmacyclics: Research Funding; IBM: Research Funding; MuTect, ABSOLTUE, MutSig and POLYSOLVER: Patents & Royalties: MuTect, ABSOLTUE, MutSig and POLYSOLVER. Shipp:BMS: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; Merck & Co.: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding; AstraZeneca: Honoraria, Membership on an entity's Board of Directors or advisory committees; Bayer: Research Funding; Gilead Sciences: Honoraria, Membership on an entity's Board of Directors or advisory committees; Takeda Pharmaceuticals: Honoraria, Membership on an entity's Board of Directors or advisory committees.

A novel soft computing method for engine RUL prediction

Prognostics is an engineering discipline focused on predicting the Remaining Useful Life (RUL) of a system or a component using raw multimedia (sensor) data. This paper presents a novel machine learning model for this task, which includes a smart ensemble of gradient boosted trees (GBT) and feed-forward neural networks. It incorporates discussions on the poor performance of MLPs and the need of ensemble models. Initial stages of data exploration and pre-processing are also comprehensively documented. Experiments are performed on the four run-to-failure C-MAPSS datasets defined by the 2008 PHM Data Challenge Competition. It concludes by presenting evaluations of multiple prediction models like MLP, SVR, CNN & gradient boosted trees (GBT). Gradient Boosted Trees are efficient in the sense that they produce an encouraging scoring model with minimum effort and also return feature importance information. The proposed method uses stacking ensemble of feed-forward neural networks and gradient boosted trees, as first level learner, and, a single-hidden layer- fully-connected neural network as the meta learner. This ensemble provides better results than any of the models alone or weighted average of their predictions. The proposed method outperforms MLP, SVR, CNN and GBT.

Multistage ensemble of feedforward neural networks for prediction of heating energy consumption

Feedforward neural network models are created for prediction of heating energy consumption of a university campus. Actual measured data are used for training and testing the models. Multistage neural network ensemble is proposed for the possible improvement of prediction accuracy. Previously trained feed-forward neural networks are first separated into clusters, using k-means algorithm, and then the best network of each cluster is chosen as a member of the ensemble. Three different averaging methods (simple, weighted and median) for obtaining ensemble output are applied. Besides this conventional approach, single radial basis neural network in the second level is used to aggregate the selected ensemble members. It is shown that heating energy consumption can be predicted with better accuracy by using ensemble of neural networks than using the best trained single neural network, while the best results are achieved with multistage ensemble.

A comparison between SARIMA models and Feed Forward Neural Network Ensemble Models for Time Series Data

In this paper, we investigate the robustness of Feed Forward Neural Network (FFNN) ensemble models applied to quarterly time series forecasting tasks, by comparing their prediction ability with that of Seasonal Auto-regressive Integrated Moving Average (SARIMA) mod- els. We obtained adequate SARIMA models which required statistical knowledge and considerable eort. On the other hand, FFNN ensemble models were readily constructed from a single FFNN template, and they produced competitive forecasts, at the level of well-constructed SARIMA models. The single template approach for adapting FFNN ensembles to multiple time series datasets can be an economic and sensible alternative if tting individual models for each time series turns out to be very time consuming. Additionally, FFNN ensembles were able to produce accurate interval estimations, in addition to good point forecasts.

Neural network studies. 4. Introduction to associative neural networks.

Associative neural network (ASNN) represents a combination of an ensemble of feed-forward neural networks and the k-nearest neighbor technique. This method uses the correlation between ensemble responses as a measure of distance amid the analyzed cases for the nearest neighbor technique. This provides an improved prediction by the bias correction of the neural network ensemble. An associative neural network has a memory that can coincide with the training set. If new data becomes available, the network further improves its predictive ability and provides a reasonable approximation of the unknown function without a need to retrain the neural network ensemble. This feature of the method dramatically improves its predictive ability over traditional neural networks and k-nearest neighbor techniques, as demonstrated using several artificial data sets and a program to predict lipophilicity of chemical compounds. Another important feature of ASNN is the possibility to interpret neural network results by analysis of correlations between data cases in the space of models. It is shown that analysis of such correlations makes it possible to provide "property-targeted" clustering of data. The possible applications and importance of ASNN in drug design and medicinal and combinatorial chemistry are discussed. The method is available on-line at http://www.vcclab.org/lab/asnn.