Commits

Show all
Author Commit Message Labels Comments Date
chuttenh
[svn r66] Fix a bug in Combiner that was misaligning genes in large data sets Add an everything flag to BNTester that will keep pairs without answers Fix a typo in Sigmoider that was flipping parameters incorrectly Fix a bug in CDatasetCompact that was mangling the everything flag
chuttenh
[svn r64] Improve Pearson correlation to ignore missing values Improve CStatistics Average and Variance to use better templating Add mean/stdev normalization to CPCL (and rearrange header a bit) Add Dat2Graph front end to the previously added CDat graph output Add PCL normalization front end to Normalize Add SQL table output to BNFunc
chuttenh
[svn r63] Fix pointer-based OpenToken to avoid running off the end of the string Add minimum correlation cutoff to MCluster/CClustHierarchical Improve PCL opening to avoid all of those awful stream/token/linebreak checks Add graph output (in several formats) to dat files
chuttenh
[svn r62] Add everything flag to Dab2Dad and compact dataset construction Fix PCL gene addition to align new features
chuttenh
[svn r61] Add default CPTs to SMILE Bayes nets for naive learning Update hierarchical clustering to MeV algorithm (much faster!) Add row expansion to full matrices/PCL files Fix MCluster to correctly deal with genes missing from the PCL Add skeleton (unimplemented) Gaussian mixture fitting to functional Bayes nets Improve text dat file opening (what was I thinking??)
chuttenh
[svn r60] Add normalization flag to Dat2Dab Fix Linux static linking for BNTester, BNConverter, BNTruster Fix Answerer to ignore empty functions Fix a typo in the functional Bayes net learner Update the SMILE Bayes net learner to smooth missing data from adjacent columns
chuttenh
[svn r59] Add exponential nodes to functional Bayes nets
chuttenh
[svn r58] Fix unnecessary (or harmful) unsigned int casts
chuttenh
[svn r57] Add beta distribution to CStatistics Update CDataset to load quantized files whenever possible Fix all of the functional Bayes net learning/evaluation stuff Add Gaussian and beta function BN learners
chuttenh
[svn r56] Safekeeping checkin before ripping apart CDataset Update CDataset and CExampleImpl to handle continuous data better Add per-node continuous/discrete checking to Bayes nets Add answer-less CDataset opening Add function-fitting Bayes net class CBayesNetFN Add CBayesNetFN access to BNConverter and BNTester
chuttenh
[svn r55] Add implementation for the -t testing argument of Answerer
chuttenh
[svn r54] Add z-score normalization to Normalizer Fix Answerer behavior under Linux Add Makefiles for BNTruster and DSLConverter Update Combiner to max rather than average Fix dat gene list Open method to not overwrite positives
chuttenh
[svn r53] Add GetGeneNames to libBioUtils CGenome Add Answerer to generate answer files from gene lists Add gene list Open method to libBioUtils CDat Update BNTruster to take the network prior into account
chuttenh
[svn r52] Add single datum evaluation to libBioUtils Bayes nets Add BNTruster to evaluate per-node BN posteriors Add DSLConverter to interconvert (X)DSL files
chuttenh
[svn r51] Add proper negative pair functional handling to DChecker and WilcoxonRankSum Fix diagonal lookup in binary half matrices
chuttenh
[svn r50] Update Combiner to take advantage of more memory (not use subsets) Update DChecker to use inverted AUC when requested Fix libBioUtils CDatasetCompact on 64-bit machines with -1s Add invertable AUC to libBioUtils CStatistics
chuttenh
[svn r49] Improve GO annotation parsing to relax and allow human annotations Add gene inclusion list to Wilcoxon rank AUC
chuttenh
[svn r48] Fix Linux compatibility in Distancer and make gene list inclusion more strict Add GetGeneNames to CGenes
chuttenh
[svn r47] Update Converter to average dats representing z-scores Add gene inclusion list processing to Distancer Fix a typo in WilcoxonRankSum statistics
chuttenh
[svn r46] Add WilcoxonRankSum (AUC) calculation to libBioUtils statistics Add extremely inefficient absolute index access to halfmatrix Fix loss of precision in normalization calculations Fix off-by-one in DChecker's calculations (use full 0-1 range)
chuttenh
[svn r45] Add masking to Dab2Dad dataset generation Improve dad SaveText output (real -1s)
chuttenh
[svn r44] Fix Dab2Dad compilation problems on Linux
chuttenh
[svn r43] Add per-Bayes net dad file creation with Dab2Dad (allowing hidden nodes) Improve dataset.cpp organization (break out CDatasetCompact*)
chuttenh
[svn r42] Add BNWrangler with functionality of BNConverter, BNTester, and DChecker Add gene filtration to IDataset instead of just compact data sets Improve CDataMask to behave more consistently and integrate with filtration Fix data sets opened on zero genes Fix a few weird size_t issues in PNL Add stream consistency checks to a few Open functions
chuttenh
[svn r41] Fix Linux shared memory code (Dab2Dad and dataset) Fix a few other Linux bugs (shared libraries, templates)
chuttenh
[svn r40] Add serializability to compact matrices and compact data sets Fix textual dat reading in a hacky way by imposing a gene limit Improve data usage in ELR learning by precalculating hidden nodes Add whole data set handling to BNConverter and BNTester Fix quantization to prevent bin overflows Update Dat2Dab to prevent weird filtering issues Add CDatasetCompactMap memory mapped compact data set Add Dab2Dad data set serialization tool
chuttenh
[svn r39] Add naive learning to Smile networks Add term gene inclusion to BNConverter/BNTester/DChecker Add normalized pearson correlation to Distancer Add z-scoring to dat file normalization
chuttenh
[svn r38] Add complete trie implementation Update ELR learner to be able to access it (but it's not actually more efficient in this case!)
chuttenh
[svn r37] Add a trie class to libBioUtils Add grouping to ELR learning to greatly improve speed Safekeeping checkin before moving over to trie grouping
chuttenh
[svn r36] Fix the ELR implementation - it might even work now!
  1. Prev
  2. Next