VertexFrame timeseries_durbin_watson_test

---

timeseries_durbin_watson_test(self, residuals)

Durbin-Watson statistics test

Parameters:

residuals : unicode

Name of the column that contains residual values

Returns:

: float64

Examples

In this example, we have a frame that contains time series values. The inspect command below shows a snippet of what the data looks like:

>>> frame.inspect()
[#]  date                      a    b              c
================================================================
[0]  2016-04-29T08:00:00.000Z   50            1.0  30.3600006104
[1]  2016-05-02T08:00:00.000Z  -50  2.09999990463  30.6100006104
[2]  2016-05-03T08:00:00.000Z   50            3.0  30.3600006104
[3]  2016-05-04T08:00:00.000Z  -50  3.90000009537  29.8500003815
[4]  2016-05-05T08:00:00.000Z   50  4.80000019073  29.8999996185
[5]  2016-05-06T08:00:00.000Z  -50            6.0  30.0400009155
[6]  2016-05-09T08:00:00.000Z   50  7.19999980927  29.7999992371
[7]  2016-05-10T08:00:00.000Z  -50            8.0  30.1399993896
[8]  2016-05-11T08:00:00.000Z   50  9.10000038147  30.0599994659
[9]  2016-05-12T08:00:00.000Z  -50  10.1999998093  29.7600002289

Calculate Durbin-Watson test statistic by giving it the name of the column that has the time series values. Let’s first calcuate the test statistic for column a:

>>> frame.timeseries_durbin_watson_test("a")
[===Job Progress===]
3.789473684210526

The test statistic close to 4 indicates negative serial correlation. Now, let’s calculate the Durbin-Watson test statistic for column b:

>>> frame.timeseries_durbin_watson_test("b")
[===Job Progress===]
0.02862014538727885

In this case, the test statistic is close to 0, which indicates positive serial correlation.