-
Notifications
You must be signed in to change notification settings - Fork 7
Expand file tree
/
Copy pathCHANGELOG.txt
More file actions
122 lines (105 loc) · 5.93 KB
/
CHANGELOG.txt
File metadata and controls
122 lines (105 loc) · 5.93 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
## [v1.0.0-beta.13] - 05.09.2024
- Fix random seed error in newer versions of MATLAB.
## [v1.0.0-beta.12-hotfix] - 21.04.2024
- Add function to check if given matrix symmetric with tolerance.
- Fix minor bug in NBSPredict_predict.
## [v1.0.0-beta.12] - 03.04.2024
- Reintroduced LOOCV for classification problems (no LOOCV for regression yet).
- Fix minor bugs in GUI.
## [v1.0.0-beta.11] - 21.08.2023
- Fix minor bugs in novel data prediction function.
- Fix minor bugs in saving results.
## [v1.0.0-beta.10] - 19.06.2023
- Refactor CPM functions.
- CPM can now provide predicted and true labels, so that user can compute further performance metrics.
- Prediction performance of combined network (positive and negative) is now provided.
- Add confound correction to CPM.
- Add permutation testing to CPM.
- Move generate_randomStream to an external function.
## [v1.0.0-beta.9] - 16.03.2022
- Drop Decision Tree Regressor and Classifier to avoid possible
incompatibilities with confound regression.
- Minor tweaks in parallel processing.
## [v1.0.0-beta.8] - 25.02.2022
- NBS-Predict is now in v1.0.0-beta.8
- Fix minor bug in feature selection (when confounds are provided in the
design matrix, but not specified in the contrast vector.).
## [v1.0.0-beta.7] - 23.02.2022
- NBS-Predict is now in v1.0.0-beta.7
- Trained model is now automatically extracted.
- Add function (NBSPredict_predict) to predict novel data using the trained model.
- Refactor several functions to make them compatible for the novel data prediction function.
- The symmetry of the connectivity matrices is now automatically checked.
- Fix minor GUI issue in the results viewer window.
- Fix edge weight issue when MAD, MSE or RMSE chosen.
- Users can also provide directory for holdout data.
- Refactor load functions to be compatible for loading a single dataset file (NodesxNodesxSubjects).
## [v1.0.0-beta.6] - 09.01.2022
- NBS-Predict is now in v1.0.0-beta.6
- User can specify number of CPU cores to use.
- Fix minor bug in Elastic Net function.
- Add progress bar for loading connectivity matrices on the command window.
- Minor change in loadData function to fasten .csv loading.
- Add balanced accuracy metric to measure accuracy in imbalanced datasets.
- Add permutation testing to compare model performance against null-distribution.
## [v1.0.0-beta.5] - 05.11.2021
- NBS-Predict is now in v1.0.0-beta.5
- NBS-Predict has now Workspace!
- GUI history has been reactivated!
- Fix bug in confusion matrix visualization.
- Fix minor bugs in GUI.
- Update user MANUAL.
- Fix compatibility issues.
## [v1.0.0-beta.4] - 28.10.2021
- NBS-Predict is now in v1.0.0-beta.4
- This version fixes the "check_classifiction function is not found" error.
## [v1.0.0-beta.3] - 23.10.2021
- NBS-Predict is now in v1.0.0-beta.3
- This version provides hotfixes for a few bugs found in v1.0.0-beta2.
- Fix compatibility issues in the NBS-Predict visualization window.
- GUI history has been deactivated until the following versions due to unexpected crashes; thus, parameters entered in the setup interface will not be saved for later use!
## [v1.0.0-beta2] - 21.10.2021
- NBS-Predict is now in v1.0.0-beta2.
- Reorganize and clarify function comments.
- Add link for the sample data.
- Update user MANUAL.
- Add wait/progress bar to the NBS-Predict visualization window.
- Minor change in identifying whether the given data is classification or regression. Target with more than 3 unique values are considered as regression problems.
- Add warnings and error messages in case no features survived the feature selection.
- Refactor a few GUI functions.
## [v1.0.0-beta1] - 13.04.2021
- NBS-Predict is now in v1.0.0-beta1.
- Add user MANUAL and data analysis tutorial.
- Fix history.mat file issue.
- Add option to set a specific random seed.
- Fix several minor issues in the GUI.
- Add functions for alternative algorithms (elastic net, lasso, p-value thresholding, top 5% and connectome-based predictive modeling (Shen et al., 2017)).
- Add simulation function to evaluate the performance of NBS-Predict and alternative algorithms on simulated network data.
- Add plot functions to plot simulation results.
## [v1.0.0-alpha3] - 06.09.2020
- NBS-Predict is now in v1.0.0-alpha3.
- Fix several bugs in the GUI.
- Fix bug in the contrast network generation algorithm.
- Fix minor bugs in assert statements.
- Minor tweaks in BrainNetViewer to visualize 3D Brain Image better.
- Add functionality to generate synthetic data for regression problems.
- Replace maximum percentile with p-value for pre-filtering.
- Changed several functions with regards to change in pre-filtering.
- Fix minor issue in function generating parameter combinations.
- divSelect, divSelectWide, simulatedAnnealing search algorithms were removed.
- Bayesian Optimization is now usable.
- Update hyperparameter space for decision tree estimator.
- Drop weight distribution plot as it might be misleading for interpretation of feature importance.
- Import circularGraph toolbox written by Paul Kassebaum to NBS-Predict to display circular network figures.
## [v1.0.0-alpha2] - 09.02.2020
- NBS-Predict is now in v1.0.0-alpha2.
- Fix several bugs in the GUI.
- Fix bug in correlation metric.
- Change default parameters for searching algorithms. Now they have more liberal parameters (e.g., reduced number of iterations).
- Add function to control confounds.
- Add a menu in GUI to choose a scaling method.
- SVM Regression is now built-in function instead of LIBSVM.
- Add Linear Regression, Logistic Regression and Linear Discriminant Analysis
- Add MinMaxScaler, StandardScaler and MaxAbsScaler
- LIBSVM is now removed. Instead, builtin SVM functions are now available to use. They give the same prediction performance as LIBSVM, but are significantly faster.
- Add preprocessing step in shrinkMat function, that features where information is less than 10% are removed. In this way, very sparse features are now removed. It might be increased to 20% in future updates.