Dataset Information

Randomly distributed embedding making short-term high-dimensional data predictable.

ABSTRACT: Future state prediction for nonlinear dynamical systems is a challenging task, particularly when only a few time series samples for high-dimensional variables are available from real-world systems. In this work, we propose a model-free framework, named randomly distributed embedding (RDE), to achieve accurate future state prediction based on short-term high-dimensional data. Specifically, from the observed data of high-dimensional variables, the RDE framework randomly generates a sufficient number of low-dimensional "nondelay embeddings" and maps each of them to a "delay embedding," which is constructed from the data of a to be predicted target variable. Any of these mappings can perform as a low-dimensional weak predictor for future state prediction, and all of such mappings generate a distribution of predicted future states. This distribution actually patches all pieces of association information from various embeddings unbiasedly or biasedly into the whole dynamics of the target variable, which after operated by appropriate estimation strategies, creates a stronger predictor for achieving prediction in a more reliable and robust form. Through applying the RDE framework to data from both representative models and real-world systems, we reveal that a high-dimension feature is no longer an obstacle but a source of information crucial to accurate prediction for short-term data, even under noise deterioration.

SUBMITTER: Ma H

PROVIDER: S-EPMC6205453 | biostudies-literature | 2018 Oct

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Randomly distributed embedding making short-term high-dimensional data predictable.

Ma Huanfei H Leng Siyang S Aihara Kazuyuki K Lin Wei W Chen Luonan L

Proceedings of the National Academy of Sciences of the United States of America 20181008 43

Future state prediction for nonlinear dynamical systems is a challenging task, particularly when only a few time series samples for high-dimensional variables are available from real-world systems. In this work, we propose a model-free framework, named randomly distributed embedding (RDE), to achieve accurate future state prediction based on short-term high-dimensional data. Specifically, from the observed data of high-dimensional variables, the RDE framework randomly generates a sufficient numb ...[more]

PMID: 30297422

Similar Datasets

Project description:Experimental techniques for measuring chromatin accessibility are expensive and time consuming, appealing for the development of computational approaches to predict open chromatin regions from DNA sequences. Along this direction, existing methods fall into two classes: one based on handcrafted k -mer features and the other based on convolutional neural networks. Although both categories have shown good performance in specific applications thus far, there still lacks a comprehensive framework to integrate useful k -mer co-occurrence information with recent advances in deep learning.We fill this gap by addressing the problem of chromatin accessibility prediction with a convolutional Long Short-Term Memory (LSTM) network with k -mer embedding. We first split DNA sequences into k -mers and pre-train k -mer embedding vectors based on the co-occurrence matrix of k -mers by using an unsupervised representation learning approach. We then construct a supervised deep learning architecture comprised of an embedding layer, three convolutional layers and a Bidirectional LSTM (BLSTM) layer for feature learning and classification. We demonstrate that our method gains high-quality fixed-length features from variable-length sequences and consistently outperforms baseline methods. We show that k -mer embedding can effectively enhance model performance by exploring different embedding strategies. We also prove the efficacy of both the convolution and the BLSTM layers by comparing two variations of the network architecture. We confirm the robustness of our model to hyper-parameters by performing sensitivity analysis. We hope our method can eventually reinforce our understanding of employing deep learning in genomic studies and shed light on research regarding mechanisms of chromatin accessibility.The source code can be downloaded from https://github.com/minxueric/ismb2017_lstm .tingchen@tsinghua.edu.cn or ruijiang@tsinghua.edu.cn.Supplementary materials are available at Bioinformatics online.

Project description:To date, much progress has been made in dietary therapy for obese patients. A low-carbohydrate diet (LCD) has reached a revival in its clinical use during the past decade with undefined mechanisms and debatable efficacy. The gut microbiota has been suggested to promote energy harvesting. Here, we propose that the gut microbiota contributes to the inconsistent outcome under an LCD. To test this hypothesis, patients with obesity or patients who were overweight were randomly assigned to a normal diet (ND) or an LCD group with ad libitum energy intake for 12 weeks. Using matched sampling, the microbiome profile at baseline and end stage was examined. The relative abundance of butyrate-producing bacteria, including Porphyromonadaceae Parabacteroides and Ruminococcaceae Oscillospira, was markedly increased after LCD intervention for 12 weeks. Moreover, within the LCD group, participants with a higher relative abundance of Bacteroidaceae Bacteroides at baseline exhibited a better response to LCD intervention and achieved greater weight loss outcomes. Nevertheless, the adoption of an artificial neural network (ANN)-based prediction model greatly surpasses a general linear model in predicting weight loss outcomes after LCD intervention. Therefore, the gut microbiota served as a positive outcome predictor and has the potential to predict weight loss outcomes after short-term LCD intervention. Gut microbiota may help to guide the clinical application of short-term LCD intervention to develop effective weight loss strategies. (This study has been registered at the China Clinical Trial Registry under approval no. ChiCTR1800015156). IMPORTANCE Obesity and its related complications pose a serious threat to human health. Short-term low-carbohydrate diet (LCD) intervention without calorie restriction has a significant weight loss effect for overweight/obese people. Furthermore, the relative abundance of Bacteroidaceae Bacteroides is a positive outcome predictor of individual weight loss after short-term LCD intervention. Moreover, leveraging on these distinct gut microbial structures at baseline, we have established a prediction model based on the artificial neural network (ANN) algorithm that could be used to estimate weight loss potential before each clinical trial (with Chinese patent number 2021104655623). This will help to guide the clinical application of short-term LCD intervention to improve weight loss strategies.

Dataset Information

Randomly distributed embedding making short-term high-dimensional data predictable.

Publications

Randomly distributed embedding making short-term high-dimensional data predictable.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets