{
  "annotation_id": "6a3227fd-8988-4518-a488-465535876b51",
  "date_created": "2026-03-02T18:01:14.815000Z",
  "date_modified": "2026-03-02T18:01:14.815000Z",
  "file_hash": "200851ca41f6875d4ae434b940f5c7db5c98cc69025a6ecfd608c2c36062145e",
  "private": false,
  "record": {
    "abstract": "A method based on wavelet transform and genetic programming is proposed for\ncharacterizing and modeling variations at multiple scales in non-stationary\ntime series. The cyclic variations, extracted by wavelets and smoothened by\ncubic splines, are well captured by genetic programming in the form of\ndynamical equations. For the purpose of illustration, we analyze two different\nnon-stationary financial time series, S\u0026P CNX Nifty closing index of the\nNational Stock Exchange (India) and Dow Jones industrial average closing values\nthrough Haar, Daubechies-4 and continuous Morlet wavelets for studying the\ncharacter of fluctuations at different scales, before modeling the cyclic\nbehavior through GP. Cyclic variations emerge at intermediate time scales and\nthe corresponding dynamical equations reveal characteristic behavior at\ndifferent scales.",
    "arxiv_id": "physics/0612221",
    "authors": [
      "Dilip P. Ahalpara",
      "Amit Verma",
      "Prasanta K. Panigrahi",
      "Jitendra C. Parikh"
    ],
    "categories": [
      "physics.data-an",
      "q-fin.ST"
    ],
    "title": "Characterizing and modeling cyclic behavior in non-stationary time   series through multi-resolution analysis",
    "url": "https://arxiv.org/abs/physics/0612221"
  },
  "schema_id": "dorsal/arxiv",
  "source": {
    "execution_id": "1db9395f-8029-4c6e-9843-9a17bc10bc0d",
    "id": "arXiv Dataset IDs",
    "type": "Model",
    "variant": "snapshot-2026-03-01",
    "version": "0.1.0"
  },
  "user_id": 1000002
}