| Name | sig-decomp JSON |
| Version |
0.3.2
JSON |
| download |
| home_page | None |
| Summary | Optimization-based signal decomposition |
| upload_time | 2024-06-07 18:31:24 |
| maintainer | None |
| docs_url | None |
| author | None |
| requires_python | >=3.6 |
| license | BSD 3-Clause License Copyright (c) 2019, Bennet Meyers All rights reserved. Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: 1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. 2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. 3. Neither the name of the copyright holder nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission. THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
| keywords |
convex optimization
optimization
|
| VCS |
 |
| bugtrack_url |
|
| requirements |
scipy
numpy
cvxpy
matplotlib
scikit-learn
qss
clarabel
pandas
|
| Travis-CI |
No Travis.
|
| coveralls test coverage |
No coveralls.
|
# OSD: Optimization(-based) Signal Decomposition
_Modeling language for finding signal decompositions_
This software provides a modeling language for describing and solving signal decomposition problems. This framework is described in detail in an acompanying [monograph](https://web.stanford.edu/~boyd/papers/sig_decomp_mprox.html), and examples are available in the [notebooks](notebooks/) directory.
New users are encouraged to try out our _no-code, no-math_ introduction to signal decomposition, available here: http://signal-decomp-tutorial.org/
## Brief overview
We formulate the problem of decomposing a signal into components as an optimization problem, where components are described by their loss functions. Once the component class loss functions are chosen, we minimize the total loss subject to replicating the given signal with the components. Our software provides a robust algorithm for carying out this decomposition, which is guaranteed to find the globally optimal descomposition when the loss functions are all convex, and is a good heuristic when they are not.
## Vector times series signals with missing entries
We consider a vector time series or signal, `y`, which possibly has missing entries. We represent the signal compactly as a `T x p` matrix, with `T` time steps and `p` entries at each time. Some of these entries may be missing—filled with `NaN` or some other missing value indicator. We will be concerned with finding decompositions that exactly match the data at the known entries.
## Signal decomposition
We model the given signal `y` as a sum (or decomposition) of `K` component, `x^1,...,x^K`. Each component `x^k` is also a `T x p` matrix, but they do not have any missing values. Indeed, we can use the values of `x^1,...,x^K` as estimates of the missing values in the original signal. This is useful for data imputation as well as model validation.
## Component classes
The `K` components are characterized by cost functions that encode the loss of or impolausibility that a component takes on a certaint value. (In some cases, we can interpret the classes statistically, with the cost function corresponding to the negative log-likelihood of some probability density function, but this is not necessary.) Our solution method is based on evaluating the _masked proximal operators_ of the class cost functions. These operators have been defined for many useful classes in the [classes](osd/classes/) module.
# Installation
We do not yet have a package released for this code, so for now, please clone the repository and set up a virtual environment with the packages listed in the [requirements file](requirements.txt).
Raw data
{
"_id": null,
"home_page": null,
"name": "sig-decomp",
"maintainer": null,
"docs_url": null,
"requires_python": ">=3.6",
"maintainer_email": "Sara Miskovich <smiskov@slac.stanford.edu>",
"keywords": "convex optimization, optimization",
"author": null,
"author_email": "Bennet Meyers <bennetm@stanford.edu>",
"download_url": "https://files.pythonhosted.org/packages/cc/a9/d0a0ccedc41dfbf7df1e2c7ee479ffa515b8741f7ca67ef83dca09047a7c/sig_decomp-0.3.2.tar.gz",
"platform": null,
"description": "# OSD: Optimization(-based) Signal Decomposition\n_Modeling language for finding signal decompositions_\n\nThis software provides a modeling language for describing and solving signal decomposition problems. This framework is described in detail in an acompanying [monograph](https://web.stanford.edu/~boyd/papers/sig_decomp_mprox.html), and examples are available in the [notebooks](notebooks/) directory.\n\nNew users are encouraged to try out our _no-code, no-math_ introduction to signal decomposition, available here: http://signal-decomp-tutorial.org/\n\n## Brief overview\n\nWe formulate the problem of decomposing a signal into components as an optimization problem, where components are described by their loss functions. Once the component class loss functions are chosen, we minimize the total loss subject to replicating the given signal with the components. Our software provides a robust algorithm for carying out this decomposition, which is guaranteed to find the globally optimal descomposition when the loss functions are all convex, and is a good heuristic when they are not.\n\n## Vector times series signals with missing entries\n\nWe consider a vector time series or signal, `y`, which possibly has missing entries. We represent the signal compactly as a `T x p` matrix, with `T` time steps and `p` entries at each time. Some of these entries may be missing\u2014filled with `NaN` or some other missing value indicator. We will be concerned with finding decompositions that exactly match the data at the known entries.\n\n## Signal decomposition\n\nWe model the given signal `y` as a sum (or decomposition) of `K` component, `x^1,...,x^K`. Each component `x^k` is also a `T x p` matrix, but they do not have any missing values. Indeed, we can use the values of `x^1,...,x^K` as estimates of the missing values in the original signal. This is useful for data imputation as well as model validation.\n\n## Component classes\n\nThe `K` components are characterized by cost functions that encode the loss of or impolausibility that a component takes on a certaint value. (In some cases, we can interpret the classes statistically, with the cost function corresponding to the negative log-likelihood of some probability density function, but this is not necessary.) Our solution method is based on evaluating the _masked proximal operators_ of the class cost functions. These operators have been defined for many useful classes in the [classes](osd/classes/) module.\n\n# Installation\n\nWe do not yet have a package released for this code, so for now, please clone the repository and set up a virtual environment with the packages listed in the [requirements file](requirements.txt).\n",
"bugtrack_url": null,
"license": "BSD 3-Clause License Copyright (c) 2019, Bennet Meyers All rights reserved. Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: 1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. 2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. 3. Neither the name of the copyright holder nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission. THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS \"AS IS\" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. ",
"summary": "Optimization-based signal decomposition",
"version": "0.3.2",
"project_urls": {
"Bug Tracker": "https://github.com/cvxgrp/signal-decomposition/issues",
"Homepage": "https://github.com/cvxgrp/signal-decomposition"
},
"split_keywords": [
"convex optimization",
" optimization"
],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "06d5e9ffaa2e58a69ba1f2cbfe31a2756e75a2209c4f8f10bb859c329f6953d6",
"md5": "05f6bd644ecc2dc316879b756c6c10eb",
"sha256": "41a018aa93be31a3983c2b8e2557f6da3d29271d8b7ee89a46452b5459462e44"
},
"downloads": -1,
"filename": "sig_decomp-0.3.2-py3-none-any.whl",
"has_sig": false,
"md5_digest": "05f6bd644ecc2dc316879b756c6c10eb",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": ">=3.6",
"size": 95759,
"upload_time": "2024-06-07T18:31:22",
"upload_time_iso_8601": "2024-06-07T18:31:22.631448Z",
"url": "https://files.pythonhosted.org/packages/06/d5/e9ffaa2e58a69ba1f2cbfe31a2756e75a2209c4f8f10bb859c329f6953d6/sig_decomp-0.3.2-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "cca9d0a0ccedc41dfbf7df1e2c7ee479ffa515b8741f7ca67ef83dca09047a7c",
"md5": "57befb6033aeae0cd0716b9327e27c40",
"sha256": "756569154377071640f295aead8bbd43b042269495b1e7404bea9b0353220f95"
},
"downloads": -1,
"filename": "sig_decomp-0.3.2.tar.gz",
"has_sig": false,
"md5_digest": "57befb6033aeae0cd0716b9327e27c40",
"packagetype": "sdist",
"python_version": "source",
"requires_python": ">=3.6",
"size": 2915164,
"upload_time": "2024-06-07T18:31:24",
"upload_time_iso_8601": "2024-06-07T18:31:24.244957Z",
"url": "https://files.pythonhosted.org/packages/cc/a9/d0a0ccedc41dfbf7df1e2c7ee479ffa515b8741f7ca67ef83dca09047a7c/sig_decomp-0.3.2.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-06-07 18:31:24",
"github": true,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"github_user": "cvxgrp",
"github_project": "signal-decomposition",
"travis_ci": false,
"coveralls": false,
"github_actions": true,
"requirements": [
{
"name": "scipy",
"specs": []
},
{
"name": "numpy",
"specs": []
},
{
"name": "cvxpy",
"specs": []
},
{
"name": "matplotlib",
"specs": []
},
{
"name": "scikit-learn",
"specs": []
},
{
"name": "qss",
"specs": []
},
{
"name": "clarabel",
"specs": []
},
{
"name": "pandas",
"specs": []
}
],
"lcname": "sig-decomp"
}
