I have not done EDA yet.
\nBut before doing this, I am wondering the time series of cryptocurrency is chaos or not?
\nIn other words, the time series of cryptocurrency has Nonlinearity or linearity?
\nSo I surveyed how to check this. And I found. It is \"Surrogate data testing.\"
\nThe deatail is written in Wikipedia.
\nHere is excerpt of the source.
\n\nSurrogate data testing[1] (or the method of surrogate data) is a statistical proof by contradiction technique and similar to parametric bootstrapping used to detect non-linearity in a time series.[2] The technique basically involves specifying a null hypothesis H0 describing a linear process and then generating several surrogate data sets according to H0 using Monte Carlo methods. A discriminating statistic is then calculated for the original time series and all the surrogate set. If the value of the statistic is significantly different for the original series than for the surrogate set, the null hypothesis is rejected and non-linearity assumed.[2]
\n
At first, I come up with ARMA model because it is very famous. But this model can be applied to LINEAR model.
\nIntuitively, this type of data maybe Nonlinear. But I don't know…
\nSo I will check it from now on on the notebook.
For details, please check this.
\nkeywords:linearlity, nonlinearity, surrogate data testing, lyapunov exponent, chaos, time-series data
", "messages": [ { "id": 1599406, "postDate": "2021-11-29T12:17:31.313Z", "content": "I have not done EDA yet.
\nBut before doing this, I am wondering the time series of cryptocurrency is chaos or not?
\nIn other words, the time series of cryptocurrency has Nonlinearity or linearity?
\nSo I surveyed how to check this. And I found. It is \"Surrogate data testing.\"
\nThe deatail is written in Wikipedia.
\nHere is excerpt of the source.
\n\nSurrogate data testing[1] (or the method of surrogate data) is a statistical proof by contradiction technique and similar to parametric bootstrapping used to detect non-linearity in a time series.[2] The technique basically involves specifying a null hypothesis H0 describing a linear process and then generating several surrogate data sets according to H0 using Monte Carlo methods. A discriminating statistic is then calculated for the original time series and all the surrogate set. If the value of the statistic is significantly different for the original series than for the surrogate set, the null hypothesis is rejected and non-linearity assumed.[2]
\n
At first, I come up with ARMA model because it is very famous. But this model can be applied to LINEAR model.
\nIntuitively, this type of data maybe Nonlinear. But I don't know…
\nSo I will check it from now on on the notebook.
For details, please check this.
\nkeywords:linearlity, nonlinearity, surrogate data testing, lyapunov exponent, chaos, time-series data
", "rawMarkdown": "I have not done EDA yet.\nBut before doing this, I am wondering the time series of cryptocurrency is **chaos** or **not**?\nIn other words, the time series of cryptocurrency has **Nonlinearity** or **linearity**?\nSo I surveyed how to check this. And I found. It is \"Surrogate data testing.\"\nThe deatail is written in [Wikipedia](https://en.wikipedia.org/wiki/Surrogate_data_testing).\nHere is excerpt of the source.\n\n> Surrogate data testing[1] (or the method of surrogate data) is a statistical proof by contradiction technique and similar to parametric bootstrapping used to detect non-linearity in a time series.[2] The technique basically involves specifying a null hypothesis H0 describing a linear process and then generating several surrogate data sets according to H0 using Monte Carlo methods. A discriminating statistic is then calculated for the original time series and all the surrogate set. If the value of the statistic is significantly different for the original series than for the surrogate set, the null hypothesis is rejected and non-linearity assumed.[2]\n\nAt first, I come up with ARMA model because it is very famous. But this model can be applied to LINEAR model.\n\nIntuitively, this type of data maybe Nonlinear. But I don't know...\nSo I will check it from now on on the notebook.\n\nFor details, please check [this](https://www.kaggle.com/c/g-research-crypto-forecasting/discussion/291478).\n\n\nkeywords:linearlity, nonlinearity, surrogate data testing, lyapunov exponent, chaos, time-series data", "votes": 1 } ], "comments": [], "raw_markdown_by_id": { "1599406": "I have not done EDA yet.\nBut before doing this, I am wondering the time series of cryptocurrency is **chaos** or **not**?\nIn other words, the time series of cryptocurrency has **Nonlinearity** or **linearity**?\nSo I surveyed how to check this. And I found. It is \"Surrogate data testing.\"\nThe deatail is written in [Wikipedia](https://en.wikipedia.org/wiki/Surrogate_data_testing).\nHere is excerpt of the source.\n\n> Surrogate data testing[1] (or the method of surrogate data) is a statistical proof by contradiction technique and similar to parametric bootstrapping used to detect non-linearity in a time series.[2] The technique basically involves specifying a null hypothesis H0 describing a linear process and then generating several surrogate data sets according to H0 using Monte Carlo methods. A discriminating statistic is then calculated for the original time series and all the surrogate set. If the value of the statistic is significantly different for the original series than for the surrogate set, the null hypothesis is rejected and non-linearity assumed.[2]\n\nAt first, I come up with ARMA model because it is very famous. But this model can be applied to LINEAR model.\n\nIntuitively, this type of data maybe Nonlinear. But I don't know...\nSo I will check it from now on on the notebook.\n\nFor details, please check [this](https://www.kaggle.com/c/g-research-crypto-forecasting/discussion/291478).\n\n\nkeywords:linearlity, nonlinearity, surrogate data testing, lyapunov exponent, chaos, time-series data" } }