Releases: brightics/studio
Releases · brightics/studio
Brightics Studio v1.1-2019.12
Release notes
Enhancement
- XGB Classification Train - Add options
- TF-IDF - Remove blank space of column names of output tables
- Latent Dirichlet Allocation - Remove blank space of column names of output tables
- Penalized Linear Regression Train - Remove prediction table from model tab
Minor Fixes
- Fix a bug when mandatory option is not entered
Brightics Studio v1.1-2019.11
Release notes
New
- Extend Datetime
- Levene's Test
- Kernel Density Estimation
- Mann Whitney Test
- Duncan Test
- Normality Test
- Ljung-Box Test
- Friedman Test
- Wilcoxon Test
- Kruskal-Wallis Test
- Transpose Time Series
- Polynomial Expansion
- String Split
- Shift Datetime
- Decompose Datetime
- Datetime Formatter
- LDA
- LDA Model
- Adaboost Regression Train
- Adaboost Regression Predict
- Isotonic Regression Train
- Isotonic Regression Predict
- MLP Regression Train
- MLP Regression Predict
- Regression Predict
- Adaboost Classification Train
- Adaboost Classification Predict
- MLP Classification Train
- MLP Classification Predict
- Classification Predict
- Spectral Clustering
- Mean Shift Clustering
- Mean Shift Clustering Predict
- Clustering Predict
- Extract Sentimental Words
- Document Summerizer (English)
- Query Executor 2
Enhancement
- Evaluate Classification - Add Average options
- Decision Tree Classification Train - Add Class Weights option
- Decision Tree Classification Predict - Add Probability Column Prefix, Suffix Type options and the probability of output
- Logistic Regression Train - Add Class Weights option
- SVM Classification Train - Add Class Weights option
- Random Forest Classification Train - Add Class Weights option
- XGB Classification Train - Add Objective for Binary Classification, Importance Type, Class Weights options
- XGB Regression Train - Add Objective, Importance Type options, Class options
- Filter - Add Start With, End With, Contain, Not Contain options
- Get Table - Shows drop down button of tables in model tab of fore function's output
- Delete all deprecated functions - Logistic Regression Train, SVM Classification Train, Pivot, Plot ROC and PR Curves, TF-IDF
- Upgrade to allow User Dictionary - Stopwords Remover, Synonym Converter, Text Search
- Delete Stopwords Remover (User Dictionary), Synonym Converter (User Dictionary), Text Search (User Dictionary
Brightics Studio v1.1-2019.09
Minor Fixes
- Support Enterprise version of UDF.
- A load function on reports has been fixed.
- Jackson-databind has been updated.
Brightics Studio v1.1-2019.08
Brightics Studio v1.1-2019.08
Brightics Studio v1.0-2019.07
Release notes
New
- Complex Chart: stacked column / 100% stacked column, area / stacked area / 100% stacked area, bubble
- Radar / Image Grid Chart has been added.
- Array To Columns
- Columns To Array
- Replace String Variable
Enhancements
- Card Chart: Allow string type, formatter option has been added.
- Line / Area Chart: A symbol sampling option has been added.
- Legend position adjustment options have been enhanced.
- Table sizing performance has been improved.
- K-Means: Remove 'legend' of figures
- Replace Missing String: Add 'Empty String as Null' option
- Read CSV: Add 'Delimiter', 'Null Filtering', 'Strip White Space in Column Name' options
- Read/Write CSV: Change the location to read/write files by using relative path
- XGB, Decision Tree, Random Forest: Add feature importance table
- XGB Regression/Classification: Change titles in Model Tab
- Logistic Regression, XGB Classification, SVM Classification: Performance tuning
- SVD: Change label name (not parameter name)
Minor Fixes
- Table border display error has been fixed.
- Table sorting initialization error has been fixed.
- A Boolean type sorting error has been fixed.
- Fixed an error that data is not quried in published reports.
- Statistic Summary bug has been fixed
- GLM Train bug has been fixed
- Add Lead Lag bug has been fixed
Brightics Studio v1.0-2019.06
Release notes
New
- Load Model (#577)
- Unload Model (#577)
- Tokenizer (Korean) (#579)
- Tokenizer (English) (#579)
- Synonym Converter (#579)
- Synonym Converter (User Dictionary) (#579)
- Stopwords Remover (#579)
- Stopwords Remover (User Dictionary) (#579)
- Word2Vec (#579)
- Word2Vec Model (#579)
- Word2Vec Similarity (#579)
- Term-document Matrix (#579)
- Term-term Matrix (#579)
- Document-Document Matrix (#579)
Enhancement
- Add UTF8 encoding to OS environment (#556)
- Random Forest Classification Train: Add the parameter class_weight (#531)
- Supports array input: Linear Regression Train, Linear Regression Predict, Logistic Regression Train, Logistic Regression Predict, Decision Tree Classification Train, Decision Tree Classification Predict, Random Forest Classification Train, Random Forest Classification Predict, SVM Classification Train, SVM Classification Predict, XGB Regression Train, XGB Regression Predict, Hierarchical Clustering, Hierarchical Clustering Post Process, K-Means, K-Means (Silhouette), K-Means Predict, Penalized Linear Regression Train, Penalized Linear Regression Predict
Minor Fixes
Brightics Studio v1.0-2019.05
Release notes
New
- Evaluate Ranking Algorithm (#515)
Enhancement
- Correlation: Add options for displaying plots (#558)
- Logistic Regression Predict: Speed up (#548)
- SVM Classification Predict: Speed up (#548)
- XGB Classification Predict: Speed up (#548)
- TF-IDF: Speed up (#546)
- Logistic Regression Train: Add statistic information such as p-value (#543)
- Transpose: Speed up (#538)
- ALS: Speed up (#501)
- Collaborative Filtering: Speed up (#501)
- Read CSV: Cancel NA filtering (#492)
- Filter: Remove single quotes in some operations (#481)
Minor Fixes
Brightics Studio v1.0-2019.04
Release notes
Enhancement
- PCA Model: Remove 'Group By' options (#511)
- Logistic Regression Train: Array of double type is supported for feature columns (#489)
- Statistic Summary: Enhance performance of group by and it applies to Statistic Summary (#488), If there is null in Group By column, error will occur but it will be fixed in the next version
- Filter: Removes single quotes in some operations (#481)
- One Way ANOVA: Add Index Column (#460)
- One Way ANOVA: Add a condition when a feature value is a not categorical value (#458)
- Linear Regression Train: VIF default option: True => False (#453)
Minor Fixes
- Add Lead Lag: Bug fix for Offset Set. Minus value is possible to enter (#503)
- Linear Regression Train: Bug fix the problem that original data enters the model, it is removed (#499)
- Evaluate Classification: Bug fix for not allowing double type for label and prediction (#484)
- Read CSV: Bug fix for encoding (#479)
- Two-Sample T Test For Stacked Data: Bug fix that hypothesized difference is printed to 0 even though it is not actually 0 in the table (#456)
Brightics Studio v1.0
Release notes
Enhancement
- Change of PostgreSQL connector from psycopg2 to pg8000 (#463)
- Random Forest Classification Train: numeric type is supported for label column (#394)
- Collaborative Filtering: add new parameters (#377)
- Add Function Column: UDF is added (#369)
- Linear Regression: performance enhancement (#367)
Minor Fixes
- Missing images in report toolbar: (#430)
- Autocorrelation: bug fix (#402)
- Hierarchical Clustering: bug fix (#400)
- Outlier Detection Tukey/Carling: bug fix (#397)
- Flatten Json: could not convert tuple, np.array, set, etc now fixed (#382)
- Bug is fixed when model has np.nan or float nan (#379)
- Collaborative Filtering: minor bug fixes (#377)
- Query Executor: Error message when empty table (#372)
- Time Series Decomposition: bug fix (#365)
Brightics Studio v1.0-beta-11
Release notes
Enhancement
- Enhance 2 functions (delete unnecessary options)