1. ClassifyR
  2. Graph
Graph of ClassifyR
  • *
  • | *
  • | *
  • * |
  • |/
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • *
  • | *
  • * |
  • | *
  • * |
  • * |
  • | *
  • * |
  • * |
  • * |
  • * |
  • | *
  • * |
  • * |
  • | *
  • | *
  • * |
  • * |
  • | *
  • * |
  • | *
  • * |
  • | *
  • * |
  • | *
  • | |\
  • | * |
  • | | *
  • | |/|
  • | |/
  • |/|
  • * |
  • * |
  • | *
  • | *
  • * |
  • | *
  • * |
  • | *
  • * |
  • | *
  • * |
  • | *
  • * |
  • |/
  • *
  • *
  • |\
  • | *
  • * |
  • |\|
  • | *
  • * |
  • |\|
  • | *
  • * |
  • |\|
  • | *
  • * |
  • |\|
  • | *
  • * |
  • |\|
  • | *
  • | *
  • | *
  • | *
  • * |
  • |\|
  • | *
  • * |
  • |\|
  • | *
  • * |
  • |\|
  • | *
  • | *
  • | *
  • | *
  • * |
  • |\|
  • | *
  • | *
  • | *
  • * |
  • |\|
  • | *
  • |/
  • *
  • |\
  • | *
  • * |
  • |\|
  • | *
  • | | *
  • * | |
  • * | |
  • * | |
  • |\| |
  • | * |
  • * | |
  • |\| |
  • | * |
  • | | *
  • * | |
  • |\| |
  • | * |
  • a974530 (HEAD -> devel, origin/master, origin/devel, origin/HEAD) - Random number sequence extraction fixed to not extract the algorithm numeric identifier and set.seed removed from package code. - Elastic net GLM keyword now uses elastic net model. New keywords defined to enable Ridge and LASSO penalties. - Added survival vignette to .Rbuildignore and missing importFrom statement for ggupset functions to address a couple of NOTEs. - Rebuilt all articles for package website. - dropSamples now consistently applied to both measurements table and outcome vector in prepareData function. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-11-11 00:30:12 +1100;
  • d59fa04 (origin/RELEASE_3_18, RELEASE_3_18) - Added survival vignette to .Rbuildignore and missing importFrom statement for ggupset functions to address a couple of NOTEs. - Rebuilt all articles for package website. - dropSamples now consistently applied to both measurements table and outcome vector in prepareData function. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-11-11 00:00:30 +1100;
  • df32d8b - Random number sequence extraction fixed to not extract the algorithm numeric identifier and set.seed removed from package code. - Elastic net GLM keyword now uses elastic net model. New keywords defined to enable Ridge and LASSO penalties. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-11-10 17:10:19 +1100;
  • cd76050 bump x.y.z version to odd y following creation of RELEASE_3_18 branch by J Wokaty <jennifer.wokaty@sph.cuny.edu> 2023-10-24 10:00:27 -0400;
  • 5bd8a9c bump x.y.z version to even y prior to creation of RELEASE_3_18 branch by J Wokaty <jennifer.wokaty@sph.cuny.edu> 2023-10-24 10:00:27 -0400;
  • 2e2b8ee - Fixed per-sample metric calculation when number of the predicted samples in cross-validation is reduced because of some error and also keep samples unable to have C-index calculated in performance vector with NA value. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-10-15 20:52:33 +1100;
  • 59705bb - Dereference facet symbols in selectionPlot. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-10-14 01:40:13 +1100;
  • f38ad64 - Got rid of most internal plots where function creates ggplot. - Update NEWS file for next version. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-10-14 00:47:29 +1100;
  • 96bac0b - In addition to categorical classes, non-censored survival events are also now equally balanced across folds to avoid creating the situation where some folds have all censored samples in them and inability to make a meaningful model. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-10-10 21:30:09 +1100;
  • a627c7e - ditribution function now automatically sorts elements into descending order and can be made to do it in ascending order. Generalised to also work with ClassifyResult that is generated for survival data and uses C-index metric instead of only working for categorical predictions. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-10-09 15:55:02 +1100;
  • 83ab19b - Omission of ggupset in DESCRIPTIitON file rectified. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-10-03 15:50:19 +1100;
  • 6cd9d3c - Always a data.frame for CoxPH training. - Discrepancy between training result and structure of pcaModel recently introduced is fixed. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-10-01 22:16:17 +1100;
  • 56eb65f - In extremeGradientBoostingPredictInterface, expand the test data set to one-hot encoding format before subsetting it by selected features from training. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-09-26 17:00:03 +1000;
  • 21ea720 - performancePlot and selectionPlot will automatically switch to UpSet plot of axis labels if multiple values are present for some labels, such as when a multiview method was used. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-09-25 14:45:02 +1000;
  • f417021 - Ranked and chosen features are recorded for dimensionality-reducing metaclassifiers PCA and prevalidation by provision of functions PCAfeatures and prevalFeatures. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-09-21 17:15:04 +1000;
  • 421356c - coxphRanking numeric subsetting corrected. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-09-21 10:30:43 +1000;
  • 528fd78 (origin/RELEASE_3_17, RELEASE_3_17) - coxphRanking numeric subsetting corrected. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-09-21 10:30:08 +1000;
  • aee895d coxphRanking fixed so that numeric variables are not ignored when there is at least one categorical variable. The bug has been present for the past year. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-09-20 21:30:17 +1000;
  • e64fee1 coxphRanking fixed so that numeric variables are not ignored when there is at least one categorical variable. The bug has been present for the past year. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-09-20 21:30:07 +1000;
  • 66982d1 - More robust handling of full model failure when some cross-validations fail, due to a subset of data being problematic. - finalModel no longer needs to be a list, but can be ANY type. More sensible since there is at most one element. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-09-19 17:05:13 +1000;
  • ac0b432 - New missing data filtering procedure for prepareData function, which chooses dropping samples if they have a higher proportion of missing values than a feature. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-09-18 13:55:07 +1000;
  • d3667ed - Missing package catching robustified. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-08-07 11:26:01 +1000;
  • 7269f60 - Missing package catching robustified. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-08-07 11:16:37 +1000;
  • ca8b2d2 Add message suggesting dependency installation by Nick <nickrrrr@gmail.com> 2023-08-02 12:49:18 +1000;
  • 665bf64 Check if classification failed by Nick <nickrrrr@gmail.com> 2023-08-02 12:48:54 +1000;
  • d106363 add edgeR check by Nick <nickrrrr@gmail.com> 2023-08-02 12:27:39 +1000;
  • 7a59b96 Survival vignette no longer hides data loading. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-07-14 14:56:43 +1000;
  • 9ec9181 Survival vignette no longer hides data loading. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-07-14 14:45:08 +1000;
  • 3df1741 Survival vignette no longer hides data loading. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-07-14 14:37:47 +1000;
  • 2cf3de2 Survival vignette no longer hides data loading. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-07-14 14:25:34 +1000;
  • d40b5a7 Survival vignette no longer hides data loading. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-07-14 13:07:00 +1000;
  • 097d0f2 Add accidentially untracked files. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-07-01 21:21:31 +1000;
  • a5692fa METABRIC clinical data and an article about survival analysis added to give an example of how to fit and evaluate survival models. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-07-01 20:45:03 +1000;
  • 5be39a2 METABRIC clinical data and an article about survival analysis added to give an example of how to fit and evaluate survival models. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-07-01 20:17:55 +1000;
  • 76205a8 - Cox net no longer has glmnet class, which messed with dispatch during prediciton. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-06-19 21:33:25 +1000;
  • 394b280 - Cox net no longer has glmnet class, which messed with dispatch during prediciton. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-06-19 21:29:18 +1000;
  • f3621e5 - In train function, only do feature selection if selectionMethod is not none. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-06-19 17:33:11 +1000;
  • 916bd75 - In train function, only do feature selection if selectionMethod is not none. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-06-19 17:19:39 +1000;
  • 3675eeb - Sample ID matching of outcome and measurements if both are named. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-06-19 12:32:12 +1000;
  • b2dbd07 Merge branch 'RELEASE_3_17' of git.bioconductor.org:packages/ClassifyR into RELEASE_3_17 by Dario Strbenac <dario@maths.usyd.edu.au> 2023-06-19 12:27:17 +1000;
  • b996a57 - Sample ID matching of outcome and measurements if both are named. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-06-19 12:25:52 +1000;
  • e92ff3d bump version by Nicholas Robertson <nickrrrr@gmail.com> 2023-06-08 15:52:09 +1000;
  • 63f34db fix logic error by Nick <nickrrrr@gmail.com> 2023-06-08 15:34:14 +1000;
  • d2da7ce Left-over command from debugging causing error removed. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-06-01 15:38:38 +1000;
  • 6e23b77 Left-over command from debugging causing error removed. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-06-01 15:35:49 +1000;
  • f32a457 - performancePlot documentation updated to mention ranking by performance. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-05-29 17:11:03 +1000;
  • bf03d77 - verbosity control in crossValidate fixed by passing verbose parameter into CV function which uses runTests. - performancePlot and .addUserLevels modifications to allow performanceAscending and performanceDescending values for ordering list elements. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-05-29 17:08:20 +1000;
  • 2d1adac - verbosity control in crossValidate fixed by passing verbose parameter into CV function which uses runTests. - performancePlot and .addUserLevels modifications to allow "performanceAscending" and "performanceDescending" values for ordering list elements. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-05-29 16:55:20 +1000;
  • 6b2eb41 - Fixes for input data preprocessing if input to crossValidate is a list of tabular data. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-05-13 21:35:43 +1000;
  • 6010330 - Fixes for input data preprocessing if input to crossValidate is a list of tabular data. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-05-13 21:32:06 +1000;
  • bf41968 - Fixed logic for automatic characteristic selection in perforancePlot and selectionPlot. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-05-10 16:46:57 +1000;
  • c1107bc - Fixed <a href="#81">#81</a> logic for automatic characteristic selection in perforancePlot and selectionPlot. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-05-10 16:40:35 +1000;
  • 987e51b - train.DataFrame checking uninitialised variables and column access fixed. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-04-27 23:11:48 +1000;
  • 4f5fb1c - train.DataFrame checking uninitialised variables and column access fixed. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-04-27 23:08:40 +1000;
  • 6348d80 bump x.y.z version to odd y following creation of RELEASE_3_17 branch by J Wokaty <jwokaty@users.noreply.github.com> 2023-04-25 10:33:41 -0400;
  • 813d1dd bump x.y.z version to even y prior to creation of RELEASE_3_17 branch by J Wokaty <jwokaty@users.noreply.github.com> 2023-04-25 10:33:41 -0400;
  • 0b7720a Merge pull request <a href="#80">#80</a> from SydneyBioX/devel by Dario Strbenac <dario.strbenac@sydney.edu.au> 2023-04-11 16:05:10 +1000;
  • 061ef9f - News file updated in preparation for next release. - verbose pass-through fixed for crissCrossValidate. - calcExternalPerformance call in crissCrossValidate fixed so that vector of predictions and known outcomes in passed in correct order. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-04-11 16:00:04 +1000;
  • 70fbb4b Merge pull request <a href="#79">#79</a> from SydneyBioX/devel by Dario Strbenac <dario.strbenac@sydney.edu.au> 2023-04-09 10:07:15 +1000;
  • 4bbe372 - Automatic choice of performance metric fixed to avoid AUC if predictions lack class scores and only are class labels. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-04-09 10:00:21 +1000;
  • 8b2e8c1 Merge pull request <a href="#78">#78</a> from SydneyBioX/devel by Dario Strbenac <dario.strbenac@sydney.edu.au> 2023-04-03 16:45:19 +1000;
  • c5db272 - Syntax warning about ..density.. deprecation from ggplot2 fixed. - calcCVperformance categorical performanceType check fixed to include the accidentally-omitted Sample Error and Sample Accuracy values. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-04-03 16:40:14 +1000;
  • 5de54d8 Merge pull request <a href="#77">#77</a> from SydneyBioX/devel by Dario Strbenac <dario.strbenac@sydney.edu.au> 2023-03-30 14:27:29 +1100;
  • 13de56f - Automatic performance calculation now selects AUC and Balanced Accuracy. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-03-30 14:26:00 +1100;
  • 75456ee Merge pull request <a href="#76">#76</a> from SydneyBioX/devel by Dario Strbenac <dario.strbenac@sydney.edu.au> 2023-03-30 13:02:04 +1100;
  • 8ca2358 - New fuction performanceTable to create a tabular summary of performances of a list of results objects. - glm family changes to quasibinomial so as to not emit warnings when weights are used for imbalanced class sizes in logistic regression classifier. - Refactoring of performance calculation functions to allow multiple performance metrics to be user-specified. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-03-30 13:00:02 +1100;
  • 617737a Merge pull request <a href="#75">#75</a> from SydneyBioX/devel by Dario Strbenac <dario.strbenac@sydney.edu.au> 2023-03-28 16:34:06 +1100;
  • 3659582 Added verbose parameter to crossValidate to quieten the messages by default but allow them to be emitted again. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-03-28 16:33:30 +1100;
  • 0f2f591 Drop indices unlisted before feature dropping in DataFrame method. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-03-23 17:18:31 +1100;
  • d791cc1 - clinicalPredictors reverted to useFeatures, but it is now a named list rather than a two-column data frame. If an assay doesn't appear in the list names, all features of that assay will be used. - If clinical data is present and useFeatures doesn't specify informative features, then no longer emit an error but a warning is instead emitted. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-03-23 17:15:06 +1100;
  • 20d29a2 Trigger GitHub pages after change of branch name. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-03-13 21:45:46 +1100;
  • 043c07f Merge pull request <a href="#73">#73</a> from SydneyBioX/master by Dario Strbenac <dario.strbenac@sydney.edu.au> 2023-03-13 21:35:22 +1100;
  • a27f89f - Input data reorganisation for GLMtrainInterface fixed. - prepareData now drops empty levels of factor and also check for matrix class correctly since matrices have two classes in recent versions of R. - selectMulti data subsetting now uses drop = FALSE to maintain data structure when selected one feature. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-03-13 21:30:07 +1100;
  • e130926 Merge pull request <a href="#71">#71</a> from SydneyBioX/master by Dario Strbenac <dario.strbenac@sydney.edu.au> 2023-02-27 16:02:45 +1100;
  • c1cc0b8 preformatted replaced by code to avoid LaTeX errors. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-02-27 16:00:51 +1100;
  • 334d70b Merge pull request <a href="#70">#70</a> from SydneyBioX/master by Dario Strbenac <dario.strbenac@sydney.edu.au> 2023-02-21 09:56:45 +1100;
  • baba87c - Fixed a couple of variables mistakes in prepareData's list method. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-02-21 09:55:28 +1100;
  • 698b7d3 Add missing alias for precisionPathwaysPredict. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-02-20 19:35:43 +1100;
  • 9387b3a - Made precision pathway evaluation functions generics. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-02-20 17:25:02 +1100;
  • 9cf4407 - Added prediction function for precusion pathway. - Added calcCostsAndPerformance, summary, bubblePlot, flowchart, and strataPlot evaluation functions for precision pathways. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-02-20 17:10:15 +1100;
  • 6fc05f0 Merge pull request <a href="#69">#69</a> from SydneyBioX/master by Dario Strbenac <dario.strbenac@sydney.edu.au> 2023-02-16 14:22:53 +1100;
  • 897acee - Elastic net GLM and ordinary GLM wrappers now automatically determine sample weights to give better performance on unbalanced classification tasks. - precisionPathwaysTrain function implemented. Prediction companion function to do in the near future. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-02-16 14:20:54 +1100;
  • 31a54d0 - Fixed bugs introduced during today's coding. - selectionMethod should be none and not NULL if to be skipped. Easier to use with naming than NULL. - elasticNetGLM made robust against choosing lambda values which cause zero features to be chosen. - elasticNetFeatures corrected to map indices back from one-hot encoding to original features. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-02-16 00:00:02 +1100;
  • 6c8638d Saving in-progess and untested changes. - useFeatures replaced by clinicalPredictors. Filtering of variables only really needed for clinical data so this simplifies the user experienece. - precisionPathyTrain function added and partially implemented. - prepareData topNvariance is now applied per-assay and has safeguards for assays with less than topNvariance features. - crossValidate S4 method signature simplification (i.e. MultiAssayExperimentOrList). by Dario Strbenac <dario@maths.usyd.edu.au> 2023-02-15 17:25:05 +1100;
  • 82808ad Merge pull request <a href="#68">#68</a> from SydneyBioX/master by Dario Strbenac <dario.strbenac@sydney.edu.au> 2023-02-13 16:02:31 +1100;
  • ca9ffe9 - measurements subsetting in crossValidate robustified so that it doesn't cause havoc when an assay has only one feature. - Ellis' bracketing in rankingDifferentMeans fixed. - CSS now correctly applied to articles, not just the vignettes. - Pkgdown style now uses Bootstrap 5 for better layouts. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-02-13 16:00:11 +1100;
  • d42901e Merge pull request <a href="#67">#67</a> from SydneyBioX/master by Dario Strbenac <dario.strbenac@sydney.edu.au> 2023-02-10 15:50:28 +1100;
  • 3edfe44 presets HTML article showing models which have tuning parameters set by default for crossValidate and all parameters which can be edited is added. - extraParams now allows removal of a parameter if the element value of the named list of is NULL. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-02-10 15:48:48 +1100;
  • 5b3d1d0 Merge pull request <a href="#66">#66</a> from SydneyBioX/master by Dario Strbenac <dario.strbenac@sydney.edu.au> 2023-02-10 13:12:46 +1100;
  • 4cad80c - extraParams parameter of crossValidate allows any parameter in a predefined parameter set (in simpleParams.R) to be overwritten or unused ones to be set by the user. - .permutations private utility function which allows any position(s) to be fixed to a certain value. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-02-10 13:10:18 +1100;
  • 04d9a3e (origin/RELEASE_3_16, RELEASE_3_16) Added Makevars to specify C++ compiler version to use. by Dario Strbenac <dario@maths.usyd.edu.au> 2023-02-03 21:20:20 +1100;
  • 7a65440 fix documentation issues causeing warnig by Nick <nickrrrr@gmail.com> 2023-02-02 10:24:44 +1100;
  • 1806ef1 Update Roxy version by Nick <nickrrrr@gmail.com> 2023-02-02 09:45:09 +1100;
  • cb13ad9 Merge pull request <a href="#63">#63</a> from SydneyBioX/master by Dario Strbenac <dario.strbenac@sydney.edu.au> 2022-12-13 10:58:25 +1100;
  • 7c9958d Corrected version number. by Dario Strbenac <dario@maths.usyd.edu.au> 2022-12-13 10:57:52 +1100;
  • b1591ab Merge pull request <a href="#62">#62</a> from SydneyBioX/master by Dario Strbenac <dario.strbenac@sydney.edu.au> 2022-12-13 10:56:48 +1100;
  • 1fbdd46 - Minor modifications to package imports to avoid notes and warnings. by Dario Strbenac <dario@maths.usyd.edu.au> 2022-12-13 10:55:11 +1100;
  • ad22f2a - selectionMethod and classifer now automatically chosen. Were already documented to be, but overlooked for code implementation. by Dario Strbenac <dario@maths.usyd.edu.au> 2022-12-08 20:44:46 +1100;
  • b056ade Merge pull request <a href="#60">#60</a> from SydneyBioX/master by Dario Strbenac <dario.strbenac@sydney.edu.au> 2022-12-08 20:35:26 +1100;
  • c6ea292 - randomSelection function added. - crissCrossValidate and crissCrossPlot contributions by Harry Robertson added and harmonised to ClassifyR code style. - selectionMethod and classifier defaults become "auto" for crossValidate. Previously only documented but not implemented in code. - predict method for standalone use now finds the correct prediction method for each trained model. by Dario Strbenac <dario@maths.usyd.edu.au> 2022-12-08 17:15:07 +1100;