Forecasting in Big Data Environments: An Adaptable and Automated Shrinkage Estimation of Neural Networks (AAShNet) View Full Text


Ontology type: schema:ScholarlyArticle      Open Access: True


Article Info

DATE

2021-11-27

AUTHORS

Ali Habibnia, Esfandiar Maasoumi

ABSTRACT

This paper considers improved forecasting in possibly nonlinear dynamic settings, with high-dimension predictors (“big data” environments). To overcome the curse of dimensionality and manage data and model complexity, we examine shrinkage estimation of a back-propagation algorithm of a neural net with skip-layer connections. We expressly include both linear and nonlinear components. This is a high-dimensional learning approach including both sparsity L1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$L_1$$\end{document} and smoothness L2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$L_2$$\end{document} penalties, allowing high-dimensionality and nonlinearity to be accommodated in one step. This approach selects significant predictors as well as the topology of the neural network. We estimate optimal values of shrinkage hyperparameters by incorporating a gradient-based optimization technique resulting in robust predictions with improved reproducibility. The latter has been an issue in some approaches. This is statistically interpretable and unravels some network structure, commonly left to a black box. An additional advantage is that the nonlinear part tends to get pruned if the underlying process is linear. In an application to forecasting equity returns, the proposed approach captures nonlinear dynamics between equities to enhance forecast performance. It offers an appreciable improvement over current univariate and multivariate models by actual portfolio performance. More... »

PAGES

363-381

References to SciGraph publications

  • 2012. Adaptive Regularization in Neural Network Modeling in NEURAL NETWORKS: TRICKS OF THE TRADE
  • 2011. Sequential Model-Based Optimization for General Algorithm Configuration in LEARNING AND INTELLIGENT OPTIMIZATION
  • 1992. Bayesian Interpolation in MAXIMUM ENTROPY AND BAYESIAN METHODS
  • 1936-09. The approximation of one matrix by another of lower rank in PSYCHOMETRIKA
  • 1986-10. Learning representations by back-propagating errors in NATURE
  • Identifiers

    URI

    http://scigraph.springernature.com/pub.10.1007/s40953-021-00275-7

    DOI

    http://dx.doi.org/10.1007/s40953-021-00275-7

    DIMENSIONS

    https://app.dimensions.ai/details/publication/pub.1143324694


    Indexing Status Check whether this publication has been indexed by Scopus and Web Of Science using the SN Indexing Status Tool
    Incoming Citations Browse incoming citations for this publication using opencitations.net

    JSON-LD is the canonical representation for SciGraph data.

    TIP: You can open this SciGraph record using an external JSON-LD service: JSON-LD Playground Google SDTT

    [
      {
        "@context": "https://springernature.github.io/scigraph/jsonld/sgcontext.json", 
        "about": [
          {
            "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/14", 
            "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
            "name": "Economics", 
            "type": "DefinedTerm"
          }, 
          {
            "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/1402", 
            "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
            "name": "Applied Economics", 
            "type": "DefinedTerm"
          }, 
          {
            "id": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/1403", 
            "inDefinedTermSet": "http://purl.org/au-research/vocabulary/anzsrc-for/2008/", 
            "name": "Econometrics", 
            "type": "DefinedTerm"
          }
        ], 
        "author": [
          {
            "affiliation": {
              "alternateName": "Department of Economics, Virginia Tech, Blacksburg, USA", 
              "id": "http://www.grid.ac/institutes/grid.438526.e", 
              "name": [
                "Department of Economics, Virginia Tech, Blacksburg, USA"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Habibnia", 
            "givenName": "Ali", 
            "id": "sg:person.013476032335.14", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.013476032335.14"
            ], 
            "type": "Person"
          }, 
          {
            "affiliation": {
              "alternateName": "Department of Economics, Emory University, Atlanta, USA", 
              "id": "http://www.grid.ac/institutes/grid.189967.8", 
              "name": [
                "Department of Economics, Emory University, Atlanta, USA"
              ], 
              "type": "Organization"
            }, 
            "familyName": "Maasoumi", 
            "givenName": "Esfandiar", 
            "id": "sg:person.013773752721.11", 
            "sameAs": [
              "https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.013773752721.11"
            ], 
            "type": "Person"
          }
        ], 
        "citation": [
          {
            "id": "sg:pub.10.1007/978-3-642-35289-8_8", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1010592940", 
              "https://doi.org/10.1007/978-3-642-35289-8_8"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/bf02288367", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1027232529", 
              "https://doi.org/10.1007/bf02288367"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-3-642-25566-3_40", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1052722418", 
              "https://doi.org/10.1007/978-3-642-25566-3_40"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1038/323533a0", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1018367015", 
              "https://doi.org/10.1038/323533a0"
            ], 
            "type": "CreativeWork"
          }, 
          {
            "id": "sg:pub.10.1007/978-94-017-2219-3_3", 
            "sameAs": [
              "https://app.dimensions.ai/details/publication/pub.1044133634", 
              "https://doi.org/10.1007/978-94-017-2219-3_3"
            ], 
            "type": "CreativeWork"
          }
        ], 
        "datePublished": "2021-11-27", 
        "datePublishedReg": "2021-11-27", 
        "description": "This paper considers improved forecasting in possibly nonlinear dynamic settings, with high-dimension predictors (\u201cbig data\u201d environments). To overcome the curse of dimensionality and manage data and model complexity, we examine shrinkage estimation of a back-propagation algorithm of a neural net with skip-layer connections. We expressly include both linear and nonlinear components. This is a high-dimensional learning approach including both sparsity L1\\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym}\n\t\t\t\t\\usepackage{amsfonts}\n\t\t\t\t\\usepackage{amssymb}\n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$L_1$$\\end{document} and smoothness L2\\documentclass[12pt]{minimal}\n\t\t\t\t\\usepackage{amsmath}\n\t\t\t\t\\usepackage{wasysym}\n\t\t\t\t\\usepackage{amsfonts}\n\t\t\t\t\\usepackage{amssymb}\n\t\t\t\t\\usepackage{amsbsy}\n\t\t\t\t\\usepackage{mathrsfs}\n\t\t\t\t\\usepackage{upgreek}\n\t\t\t\t\\setlength{\\oddsidemargin}{-69pt}\n\t\t\t\t\\begin{document}$$L_2$$\\end{document} penalties, allowing high-dimensionality and nonlinearity to be accommodated in one step. This approach selects significant predictors as well as the topology of the neural network. We estimate optimal values of shrinkage hyperparameters by incorporating a gradient-based optimization technique resulting in robust predictions with improved reproducibility. The latter has been an issue in some approaches. This is statistically interpretable and unravels some network structure, commonly left to a black box. An additional advantage is that the nonlinear part tends to get pruned if the underlying process is linear. In an application to forecasting equity returns, the proposed approach captures nonlinear dynamics between equities to enhance forecast performance. It offers an appreciable improvement over current univariate and multivariate models by actual portfolio performance.", 
        "genre": "article", 
        "id": "sg:pub.10.1007/s40953-021-00275-7", 
        "inLanguage": "en", 
        "isAccessibleForFree": true, 
        "isPartOf": [
          {
            "id": "sg:journal.1027984", 
            "issn": [
              "0971-1554", 
              "2364-1045"
            ], 
            "name": "Journal of Quantitative Economics", 
            "publisher": "Springer Nature", 
            "type": "Periodical"
          }, 
          {
            "issueNumber": "Suppl 1", 
            "type": "PublicationIssue"
          }, 
          {
            "type": "PublicationVolume", 
            "volumeNumber": "19"
          }
        ], 
        "keywords": [
          "shrinkage estimation", 
          "gradient-based optimization techniques", 
          "curse of dimensionality", 
          "nonlinear part", 
          "optimization techniques", 
          "model complexity", 
          "nonlinear components", 
          "neural network", 
          "optimal value", 
          "skip-layer connections", 
          "back-propagation algorithm", 
          "robust predictions", 
          "approach captures", 
          "network structure", 
          "underlying process", 
          "estimation", 
          "neural nets", 
          "forecast performance", 
          "nonlinearity", 
          "hyperparameters", 
          "sparsity", 
          "dynamic setting", 
          "dimensionality", 
          "smoothness", 
          "black box", 
          "portfolio performance", 
          "topology", 
          "curse", 
          "approach", 
          "network", 
          "algorithm", 
          "dynamics", 
          "additional advantage", 
          "learning approach", 
          "complexity", 
          "performance", 
          "model", 
          "prediction", 
          "appreciable improvement", 
          "forecasting", 
          "applications", 
          "penalty", 
          "connection", 
          "data environment", 
          "technique", 
          "big data environment", 
          "structure", 
          "advantages", 
          "nets", 
          "step", 
          "values", 
          "equity returns", 
          "data", 
          "multivariate model", 
          "box", 
          "process", 
          "components", 
          "part", 
          "improved reproducibility", 
          "issues", 
          "improvement", 
          "setting", 
          "capture", 
          "univariate", 
          "environment", 
          "return", 
          "predictors", 
          "reproducibility", 
          "equity", 
          "manage data", 
          "significant predictors", 
          "paper"
        ], 
        "name": "Forecasting in Big Data Environments: An Adaptable and Automated Shrinkage Estimation of Neural Networks (AAShNet)", 
        "pagination": "363-381", 
        "productId": [
          {
            "name": "dimensions_id", 
            "type": "PropertyValue", 
            "value": [
              "pub.1143324694"
            ]
          }, 
          {
            "name": "doi", 
            "type": "PropertyValue", 
            "value": [
              "10.1007/s40953-021-00275-7"
            ]
          }
        ], 
        "sameAs": [
          "https://doi.org/10.1007/s40953-021-00275-7", 
          "https://app.dimensions.ai/details/publication/pub.1143324694"
        ], 
        "sdDataset": "articles", 
        "sdDatePublished": "2022-05-10T10:26", 
        "sdLicense": "https://scigraph.springernature.com/explorer/license/", 
        "sdPublisher": {
          "name": "Springer Nature - SN SciGraph project", 
          "type": "Organization"
        }, 
        "sdSource": "s3://com-springernature-scigraph/baseset/20220509/entities/gbq_results/article/article_878.jsonl", 
        "type": "ScholarlyArticle", 
        "url": "https://doi.org/10.1007/s40953-021-00275-7"
      }
    ]
     

    Download the RDF metadata as:  json-ld nt turtle xml License info

    HOW TO GET THIS DATA PROGRAMMATICALLY:

    JSON-LD is a popular format for linked data which is fully compatible with JSON.

    curl -H 'Accept: application/ld+json' 'https://scigraph.springernature.com/pub.10.1007/s40953-021-00275-7'

    N-Triples is a line-based linked data format ideal for batch operations.

    curl -H 'Accept: application/n-triples' 'https://scigraph.springernature.com/pub.10.1007/s40953-021-00275-7'

    Turtle is a human-readable linked data format.

    curl -H 'Accept: text/turtle' 'https://scigraph.springernature.com/pub.10.1007/s40953-021-00275-7'

    RDF/XML is a standard XML format for linked data.

    curl -H 'Accept: application/rdf+xml' 'https://scigraph.springernature.com/pub.10.1007/s40953-021-00275-7'


     

    This table displays all metadata directly associated to this object as RDF triples.

    164 TRIPLES      22 PREDICATES      103 URIs      89 LITERALS      6 BLANK NODES

    Subject Predicate Object
    1 sg:pub.10.1007/s40953-021-00275-7 schema:about anzsrc-for:14
    2 anzsrc-for:1402
    3 anzsrc-for:1403
    4 schema:author Nbcb4666219294f04af37623f76e9d2fa
    5 schema:citation sg:pub.10.1007/978-3-642-25566-3_40
    6 sg:pub.10.1007/978-3-642-35289-8_8
    7 sg:pub.10.1007/978-94-017-2219-3_3
    8 sg:pub.10.1007/bf02288367
    9 sg:pub.10.1038/323533a0
    10 schema:datePublished 2021-11-27
    11 schema:datePublishedReg 2021-11-27
    12 schema:description This paper considers improved forecasting in possibly nonlinear dynamic settings, with high-dimension predictors (“big data” environments). To overcome the curse of dimensionality and manage data and model complexity, we examine shrinkage estimation of a back-propagation algorithm of a neural net with skip-layer connections. We expressly include both linear and nonlinear components. This is a high-dimensional learning approach including both sparsity L1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$L_1$$\end{document} and smoothness L2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$L_2$$\end{document} penalties, allowing high-dimensionality and nonlinearity to be accommodated in one step. This approach selects significant predictors as well as the topology of the neural network. We estimate optimal values of shrinkage hyperparameters by incorporating a gradient-based optimization technique resulting in robust predictions with improved reproducibility. The latter has been an issue in some approaches. This is statistically interpretable and unravels some network structure, commonly left to a black box. An additional advantage is that the nonlinear part tends to get pruned if the underlying process is linear. In an application to forecasting equity returns, the proposed approach captures nonlinear dynamics between equities to enhance forecast performance. It offers an appreciable improvement over current univariate and multivariate models by actual portfolio performance.
    13 schema:genre article
    14 schema:inLanguage en
    15 schema:isAccessibleForFree true
    16 schema:isPartOf N2e387790061d4b1486a7cd760be15aa2
    17 N61735e86735b490daed4830371a31f3a
    18 sg:journal.1027984
    19 schema:keywords additional advantage
    20 advantages
    21 algorithm
    22 applications
    23 appreciable improvement
    24 approach
    25 approach captures
    26 back-propagation algorithm
    27 big data environment
    28 black box
    29 box
    30 capture
    31 complexity
    32 components
    33 connection
    34 curse
    35 curse of dimensionality
    36 data
    37 data environment
    38 dimensionality
    39 dynamic setting
    40 dynamics
    41 environment
    42 equity
    43 equity returns
    44 estimation
    45 forecast performance
    46 forecasting
    47 gradient-based optimization techniques
    48 hyperparameters
    49 improved reproducibility
    50 improvement
    51 issues
    52 learning approach
    53 manage data
    54 model
    55 model complexity
    56 multivariate model
    57 nets
    58 network
    59 network structure
    60 neural nets
    61 neural network
    62 nonlinear components
    63 nonlinear part
    64 nonlinearity
    65 optimal value
    66 optimization techniques
    67 paper
    68 part
    69 penalty
    70 performance
    71 portfolio performance
    72 prediction
    73 predictors
    74 process
    75 reproducibility
    76 return
    77 robust predictions
    78 setting
    79 shrinkage estimation
    80 significant predictors
    81 skip-layer connections
    82 smoothness
    83 sparsity
    84 step
    85 structure
    86 technique
    87 topology
    88 underlying process
    89 univariate
    90 values
    91 schema:name Forecasting in Big Data Environments: An Adaptable and Automated Shrinkage Estimation of Neural Networks (AAShNet)
    92 schema:pagination 363-381
    93 schema:productId N0ab9db582d6241be9c6639dd10f5a18a
    94 Nd6be58ff04ce4963816c28d5d844d5b6
    95 schema:sameAs https://app.dimensions.ai/details/publication/pub.1143324694
    96 https://doi.org/10.1007/s40953-021-00275-7
    97 schema:sdDatePublished 2022-05-10T10:26
    98 schema:sdLicense https://scigraph.springernature.com/explorer/license/
    99 schema:sdPublisher Ndfaa1886eca94751b0a54a674644330a
    100 schema:url https://doi.org/10.1007/s40953-021-00275-7
    101 sgo:license sg:explorer/license/
    102 sgo:sdDataset articles
    103 rdf:type schema:ScholarlyArticle
    104 N0ab9db582d6241be9c6639dd10f5a18a schema:name doi
    105 schema:value 10.1007/s40953-021-00275-7
    106 rdf:type schema:PropertyValue
    107 N2e387790061d4b1486a7cd760be15aa2 schema:volumeNumber 19
    108 rdf:type schema:PublicationVolume
    109 N37c4b0cbd76f47908598be907ab2cb40 rdf:first sg:person.013773752721.11
    110 rdf:rest rdf:nil
    111 N61735e86735b490daed4830371a31f3a schema:issueNumber Suppl 1
    112 rdf:type schema:PublicationIssue
    113 Nbcb4666219294f04af37623f76e9d2fa rdf:first sg:person.013476032335.14
    114 rdf:rest N37c4b0cbd76f47908598be907ab2cb40
    115 Nd6be58ff04ce4963816c28d5d844d5b6 schema:name dimensions_id
    116 schema:value pub.1143324694
    117 rdf:type schema:PropertyValue
    118 Ndfaa1886eca94751b0a54a674644330a schema:name Springer Nature - SN SciGraph project
    119 rdf:type schema:Organization
    120 anzsrc-for:14 schema:inDefinedTermSet anzsrc-for:
    121 schema:name Economics
    122 rdf:type schema:DefinedTerm
    123 anzsrc-for:1402 schema:inDefinedTermSet anzsrc-for:
    124 schema:name Applied Economics
    125 rdf:type schema:DefinedTerm
    126 anzsrc-for:1403 schema:inDefinedTermSet anzsrc-for:
    127 schema:name Econometrics
    128 rdf:type schema:DefinedTerm
    129 sg:journal.1027984 schema:issn 0971-1554
    130 2364-1045
    131 schema:name Journal of Quantitative Economics
    132 schema:publisher Springer Nature
    133 rdf:type schema:Periodical
    134 sg:person.013476032335.14 schema:affiliation grid-institutes:grid.438526.e
    135 schema:familyName Habibnia
    136 schema:givenName Ali
    137 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.013476032335.14
    138 rdf:type schema:Person
    139 sg:person.013773752721.11 schema:affiliation grid-institutes:grid.189967.8
    140 schema:familyName Maasoumi
    141 schema:givenName Esfandiar
    142 schema:sameAs https://app.dimensions.ai/discover/publication?and_facet_researcher=ur.013773752721.11
    143 rdf:type schema:Person
    144 sg:pub.10.1007/978-3-642-25566-3_40 schema:sameAs https://app.dimensions.ai/details/publication/pub.1052722418
    145 https://doi.org/10.1007/978-3-642-25566-3_40
    146 rdf:type schema:CreativeWork
    147 sg:pub.10.1007/978-3-642-35289-8_8 schema:sameAs https://app.dimensions.ai/details/publication/pub.1010592940
    148 https://doi.org/10.1007/978-3-642-35289-8_8
    149 rdf:type schema:CreativeWork
    150 sg:pub.10.1007/978-94-017-2219-3_3 schema:sameAs https://app.dimensions.ai/details/publication/pub.1044133634
    151 https://doi.org/10.1007/978-94-017-2219-3_3
    152 rdf:type schema:CreativeWork
    153 sg:pub.10.1007/bf02288367 schema:sameAs https://app.dimensions.ai/details/publication/pub.1027232529
    154 https://doi.org/10.1007/bf02288367
    155 rdf:type schema:CreativeWork
    156 sg:pub.10.1038/323533a0 schema:sameAs https://app.dimensions.ai/details/publication/pub.1018367015
    157 https://doi.org/10.1038/323533a0
    158 rdf:type schema:CreativeWork
    159 grid-institutes:grid.189967.8 schema:alternateName Department of Economics, Emory University, Atlanta, USA
    160 schema:name Department of Economics, Emory University, Atlanta, USA
    161 rdf:type schema:Organization
    162 grid-institutes:grid.438526.e schema:alternateName Department of Economics, Virginia Tech, Blacksburg, USA
    163 schema:name Department of Economics, Virginia Tech, Blacksburg, USA
    164 rdf:type schema:Organization
     




    Preview window. Press ESC to close (or click here)


    ...