Releases: IBM/TSML.jl
Releases · IBM/TSML.jl
v2.3.0
Refactor for pure julia codes; TSMLextra for external libs
- refactor code for fast docker image
- refactor code for package compiler dynamic library support
- separate RCall and PyCall support into TSMLextra
v2.1.0 (2019-05-29)
Closed issues:
- Parameterise aggregator function for DateValgator and DateValNNer (#51)
- Error tagging new release (#40)
- Version 2.0.0 Release (#39)
- Documentation of main workflow with Documenter (#30)
- Outlierlizer: Normalize outliers (#19)
- Default TS cleaning/normalization pipeline as one function for CLI (#14)
Merged pull requests:
Removed deps to R and Python, full tutorial and documentation
Commits
- Add website link for documentation (5292115)
- DateValNNer manual doc (3a93648)
- Documentation and Tutorial (425751c)
- Merge branch 'master' of https://github.com/IBM/TSML.jl (f0b1785)
- Merge branch 'master' of https://github.com/IBM/TSML.jl (8d231ba)
- Set theme jekyll-theme-midnight (3f2614b)
- Set theme jekyll-theme-minimal (fa168ee)
- Update README.md (58dfeff)
- Update README.md (80030b9)
- Update ROADMAP.md (e1be376)
- add Daily support in DateValizer; manual for imputation (1e18417)
- add TSML to docs Project.toml (5c108d4)
- add description of the stored model (3b6204f)
- add documenter support (4a74e7c)
- add gitignore files (8408a7c)
- add label to the Matrifier output column (d8fc1c9)
- adding Manifest to gitignore (0646a81)
- adding aggregators, pipeline, stat, and detectors docs (2265a8e)
- adding build docs (d3d27cc)
- adding compat of julia 1 at Project.toml (9176388)
- adding coverage (e824e30)
- adding coverage and documentation in travis (07f072e)
- adding deps to Project.toml (19c51f3)
- adding docs for Value Matrification (8751d7b)
- adding docstring to decision tree code (f84c401)
- adding more demo samples in notebook tutorial (34c8b9a)
- adding more docs; monotonic doc; ts detection (c763b56)
- adding pipeline demo examples (d16ae7f)
- adding preprocessing of dates and values (ac80b57)
- adding statifier demo examples (95c0e6f)
- change capitalization of statifier output; remove unnecessary man docs (8ee30ff)
- change matrix to dataframes (05c1a4d)
- cleaning up manifest files (4f7655c)
- cleanup testdata_output (c5bd6c1)
- combine workflow for matrifier and dateifier (7659a37)
- create index.html for the tsml project (3401397)
- doc link update (654250c)
- docs for TS classification (9714495)
- documentation in pdf (93664c4)
- documentation link (ce62685)
- edits of manual docs (99b03bd)
- fix bug for julia 1.1; update make.jl (1bc63a0)
- fix docs (48f0778)
- hide testdata_output.csv from git (a8c968e)
- include more useful help links (09015e9)
- initial documentation (4d895ae)
- manual for aggregation doc (94f3bab)
- merge master (69e141c)
- minor edit (2fafc89)
- minor edit of overview (60b61ff)
- minor edits (73e9a5b)
- more readme updates (83f4722)
- place cli wrapper in the backlog (5e0dac9)
- refresh index.html to point to docs (af149fe)
- remove PyCall and RCall deps (689e210)
- remove and fix outline (6c84e73)
- remove block stat in tsclassifier output (f429e65)
- remove default block stat in tsclassifier output (df941bb)
- remove docs/build from git view (1a58e25)
- remove index.html redirect as it's not needed after gh_pages support (cfa1ba5)
- remove manifest file (29707ae)
- remove python 2 check (00de0c5)
- removing manual doc/build dir; let travis generate docs (69eff6e)
- reorganize documentation/docker files (72740c5)
- trying another doc config in travis (0254378)
- update README (9196b21)
- update README (c016184)
- update Readme for the link of doc and installation (4eb936a)
- update index.html (cec4d72)
- update link (4261e31)
- update link from travis generated docs (774dd44)
- update travis for julia 1.1 support (4d22023)
- value and date processing subsections (a3ddbb7)
Merged PRs
This changelog was automatically generated, and might contain inaccuracies.
TS auto classification, monotonic data detection/normalization
- Statifier <: Transformer - scalar stats for data quality characterization
- MonotonicFilter <: Transformer - convert monotonic data using finite difference operator
- TSClassifier <: Transformer - automatic classification of TS data type
- High-level wrapper for CLI automation and interfacing with other programs in the docker/shell
- Dockerized branch for Kubernetes deployment
Ensemble wrappers, multiformat data readers/writers
- RandomForest <: TSLearner - RF regression/classification wrapper
- PrunedTree <: TSLearner - decision tree regression/classification wrapper
- Adaboost <: TSLearner - Adaboost regression/classification wrapper
- DataReader <: Transformer - hdf5/feather/jld/csv multiformat reader
- DataWriter <: Transformer - hdf5/feather/jld/csv multiformat writer
Docker and Notebook Support
- Dockerization - dockerized notebook tutorial and dockerized TSML
Matrify TS for ML workflow
- Matrifier <: Transformer - transform vector of values into matrix by sliding windows
- Dateifier <: Transformer - get the date boundaries in the sliding windows to correspond with matrifier output
- DateValNNer <: Transformer - nearest neighbor replacement of missing data
- CSVDateValReader <: Transformer - CSV reader
- CSVDateValWriter <: Transformer - CSV writer
Initial Release: Base data structures and ML wrappers
- Transformer - abstract class with fit and transform interfaces to be overloaded
- TSLearner <: Transformer - learners for classification/prediction with fit function for training and transform for prediction
- Baseline <: TSLearner - returns the mode for classification and usually provides the worst case result
- CaretLearner <: TSLearner - API wrapper to expose caret regression/classification libs
- SKLearner <: TSLearner - API wrapper to expose scikitlearn regression/classification libs
Identity <: Transformer - identity learner (returns mirror image) - Imputer <: Transformer - removes missing values
- Pipeline <: Transformer - iteratively calls fit! and transform! to the set of transformers in the workflow
- DateValizer <: Transformer - replace missings with medians grouped by datetime period
- DateValgator <: Transformer - Aggregate values grouped by datetime period