@@ -6,8 +6,43 @@ The PyMFE releases are available in PyPI_ and GitHub_.
6
6
.. _GitHub : https://github.com/ealcobaca/pymfe/releases
7
7
8
8
9
- Version 0.2.0
10
- -------------
9
+ Version 0.3.0 (Available on PyPI)
10
+ ---------------------------------
11
+ * Metafeature extraction with confidence intervals
12
+
13
+ * Pydoc fixes and package documentation/code consistency improvements
14
+
15
+ * Reformatted 'model-based' group metafeature extraction methods arguments to
16
+ a consistent format (all model-based metafeatures now receive a single
17
+ mandatory argument 'dt_model', and all other arguments are optional
18
+ arguments from precomputations.) Now it is much easier to use those
19
+ methods directly without the main class (mfe) filter, if desired.
20
+
21
+ * Now accepting user custom arguments in precomputation methods.
22
+
23
+ * Added 'extract_from_model' MFE method, making easy to extract model-based
24
+ metafeatures from a pre-fitted model without using the training data.
25
+
26
+ * Memory issues
27
+
28
+ * Now handling memory errors in precomputations, postcomputations and
29
+ metafeature extraction as a regular exception.
30
+
31
+ * Categorical attributes one-hot encoding option
32
+
33
+ * Added option to encode categorical attributes using one-hot encoding
34
+ instead of the current gray encoding.
35
+
36
+ * New nan-resilient summary functions
37
+
38
+ * All summary functions now can be calculated ignoring 'nan' values, using
39
+ its nan-resilient version.
40
+
41
+ * Online documentation improvement
42
+
43
+
44
+ Version 0.2.0 (Available on PyPI)
45
+ ---------------------------------
11
46
* New meta-feature groups
12
47
13
48
* Complexity
@@ -35,8 +70,8 @@ Version 0.2.0
35
70
* Statistical group updated
36
71
37
72
38
- Version 0.1.1
39
- -------------
73
+ Version 0.1.1 (Available on PyPI)
74
+ ---------------------------------
40
75
* Bugs solved
41
76
42
77
* False positive of mypy fixed
0 commit comments