{
  "abstract": "Forecasting multiple time-series with high-dimensional covariates presents a core challenge: unifying common temporal patterns while retaining meaningful series-specific information. We introduce Hopformer (Homogeneity-Pursuit Transformer), a two-stage framework that addresses this challenge. In the first stage, we perform a Sparsity Pattern Aggregation (SPA) scheme extracting a common low-variance trend that incorporates the covariates. This acts as a homogenization layer. In the second stage, a LoRA-fine-tuned Transformer models the remaining complex dependencies in the residual. Our method is theoretically grounded. We prove that SPA achieves a near-optimal bias-variance trade-off via an oracle inequality. We also provide generalization bounds for the second stage under dependent time series data. Hopformer sets a new state of the art, improving MASE by an average of 6.56% across synthetic and real-world forecasting benchmarks.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2607.22299",
  "author_total": 6,
  "authors": [
    {
      "name": "Wan Zhang"
    },
    {
      "name": "Qinjie Lin"
    },
    {
      "name": "Chan Lee"
    },
    {
      "name": "Weijian Li"
    },
    {
      "name": "Han Liu"
    },
    {
      "name": "Kai Zhang"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/9ac0qhnvx0me3.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2607.22299",
  "fulltext_url": null,
  "id": "9ac0qhnvx0me3",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2607.22299",
  "language": null,
  "licence": "CC-BY-4.0",
  "licence_label": "CC BY 4.0",
  "licence_source": "article",
  "licence_url": "http://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": "CC0-1.0",
  "native_subjects": [
    [
      "arxiv",
      "cs.LG"
    ],
    [
      "arxiv",
      "stat.ML"
    ],
    [
      "openalex",
      "1711|Time Series Analysis and Forecasting"
    ]
  ],
  "openalex": "W7171401886",
  "pdf_url": "https://arxiv.org/pdf/2607.22299",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-13",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/9ac0qhnvx0me3.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:29:42Z",
      "native_id": "2607.22299",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2607.22299"
    }
  ],
  "subjects": [
    {
      "field": "mathematics",
      "origin": "arxiv:stat.ML",
      "subfield": "statistics"
    }
  ],
  "title": "Hopformer: Homogeneity-Pursuit Transformer for Time Series Forecasting",
  "title_text": "Hopformer: Homogeneity-Pursuit Transformer for Time Series Forecasting",
  "type": "preprint",
  "updated_at": "2026-10-30T12:29:42Z",
  "url": "https://research.draconitic.com/paper/9ac0qhnvx0me3",
  "venue": "arXiv",
  "versions": []
}