Datasets¶
datasets ¶
Data loaders. The chapter 4 data set is shipped inside the package (qlpca/data).
data_dir ¶
data_dir()
Folder that holds the data files.
Returns:
| Type | Description |
|---|---|
Path
|
The folder given by the |
load_st_gender ¶
load_st_gender()
Science & Technology gender data, 15 EU countries (thesis chapter 4, Appendix A.1).
The file is shipped with the package and read with :mod:importlib.resources, so this works after
pip install. Set the environment variable QLPCA_DATA to a folder containing
st_gender_eu15.csv to read another copy. The data are not covered by the MIT licence of the code
(see qlpca/data/README.md).
Returns:
| Name | Type | Description |
|---|---|---|
X |
ndarray of shape (15, 11)
|
The 11 indicators. |
variables |
list of str
|
Variable names. |
countries |
list of str
|
Country codes. |
make_cylinder ¶
make_cylinder(n=51, noise=0.0, random_state=None)
Simulated "cylinder problem" (thesis section 3.3).
Twelve variables derived from the height a and base area b of n cylinders, with a, b ~ U(0, 1) (stated
in the MSc thesis, chapter 6). With noise, every variable is mapped to its linearising scale (log;
log(tan) for the arctan variable and log(cot) for the arccot variable), normal noise with standard
deviation noise times the sample standard deviation of the variable on that scale is added, and
the result is mapped back (docs/SPEC.md sections 7 and 9).
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
n
|
int
|
Number of cylinders (thesis: 51). |
51
|
noise
|
float
|
0, 0.10 or 0.25 in the thesis. |
0.0
|
random_state
|
(None, int or Generator)
|
Seed. |
None
|
Returns:
| Name | Type | Description |
|---|---|---|
X |
ndarray of shape (n, 12)
|
The simulated variables. |
names |
list of str
|
Variable names. |