Downloads:
Beta v3 (March 2020)
This version is not entirely documented or tested.
Includes many updates:
Includes many updates:
- All core algorithms now functions with missing data which is represented by NaN's in the data matrix
- RCSVLoad() will read CSV files with blanks as missing and convert them to NaN's
- RCSVLoad() will read CSV files with blanks as missing and convert them to NaN's
- Improved Bootstrapping, including integration with CV
- Bootstrap Balancing of phenotypes that are heavily skewed to one value or the other
- Tree-Augmented Naive Bayes models for Discrete or Continuous networks
- Minor bug fixes
| cgbayesnets_3_23_20.zip |
Beta Version: (beta v2)
This version has not been entirely tested.
Includes updates for :
Includes updates for :
- predicting continuous phenotypes using expected value conditioned on evidence
- entering values for a subset of nodes in the BN and computing most likely outcomes for remaining nodes
- computation of posterior probability of observed evidence for all nodes
- support for simple Two-Stage Dynamic BNs for modeling time-series data.
- Naive Bayes models with sparse matrices for large datasets
| cgbayesnets_9_7_15.zip |
Current Version:
We suggest using the most updated version.
This version is packaged w/o the GEO dataset used in the PLOS Comp Bio paper by McGeachie, Chang, and Weiss.
This version is packaged w/o the GEO dataset used in the PLOS Comp Bio paper by McGeachie, Chang, and Weiss.
| cgbayesnets_7_14_14.zip |
Published Version matching PLOS Comp Bio paper:
Version 2.0! Large overhaul to make code more uniform and have computations in terms of the BayesNet class object.
Includes full GEO dataset for use with examples in supplementary data
Includes full GEO dataset for use with examples in supplementary data
| cgbayesnets_2_20_14.zip |
Development Versions. WARNING!! Bugs present!!