New
| Function | What's new! | | ------------------- | ------------------------------- | | coeff | Coefficient matrix. |
Improved (by speed, correctness or options)
| Function | What's new! | | ------------------- | ------------------------------- | | cor.fsreg | Bug fix in the adjusted R-squared computation. | | ompr | Bug fix in the adjusted R-squared computation. | | rmdp | Support parallel. | | colMads | Supports data.frame | | eachrow | Fix bug. | | colMedians | Fix bug about data.frame. | | rowMedians | Fix bug. | | colVars | Fix bug about data.frame. | | Dist(...,method = "minimum") | Bug fix. | | Dist(...,method = "bhattacharyya") | Bug fix . | | Dist(...,method = "itakura_saito") | Bug fix. Supports parallel. | | Dist(...,method = "canberra1") | Depricated. | | Dist(...,method = "canberra2") | Depricated. | | Dist(...,method = "jensen_shannon") | Supports parallel. | | Dist(...,method = "kullback_leibler") | Supports parallel. | | Dist | New methods ("canberra","chi_square","soergel","sorensen","cosine","wave_hedges","motyka","harmonic_mean","jeffries_matusita","gower","kulczynski") | | dista | All the methods from the Dist function are now supported. | | total.dist | All the methods from the Dist function are now supported. | | total.dista | All the methods from the Dist function are now supported. | | Sort | Supports parallel. Available only in Windows and most unix. | | Order | Supports parallel. Available only in Windows and most unix. | | Rank | Supports parallel. Available only in Windows and most unix. Add extra option for stable algorithm. | | checkUsage | checking if usage lines are wider than 90 characters. |
LinkingTo (by speed, correctness or options)
| Function/Structure | What's new! | | ------------------- | ------------------------------- | | colMads | Supports data.frame |
Improved (by speed, correctness or options)
| Function | What's new! | | ------------------- | ------------------------------- | | colMads | Supports data.frame | | colShuffle | Supports data.frame | | colMedian | Supports data.frame | | colRanks | Supports data.frame and fix bug about method "min". | | colSort | Supports data.frame | | colMaxs | Supports data.frame | | colMins | Supports data.frame | | colMinsMaxs | Supports data.frame and add extra option for parallelism. | | colrange | Supports data.frame and add extra option for parallelism. | | checkExamples | Add extra option for running examples inside environment. Treat special functions with new export attribute. | | sort_mat | Depricated. |
LinkingTo (by speed, correctness or options)
| Function/Structure | What's new! | | ------------------- | ------------------------------- | | colMads | Supports data.frame | | colShuffle | Supports data.frame | | colMedian | Supports data.frame | | colSort | Supports data.frame | | colMaxs | Supports data.frame | | colMins | Supports data.frame | | colMinsMaxs | Supports data.frame and add extra option for parallelism. | | NA | A structure for manage R's NA values. Supports {int, double, bool, string}. | | Type | A structure for manage R's types. Supports {REAL,INT,CHAR,FACTOR,LIST,DATAFRAME}. |
Comments
From now on all the function will be inside namespace Rfast. Sub-namespaces "vector" and "matrix" have been removed.
Improved (by speed, correctness or options)
| Function | What's new! | | ------------------- | ------------------------------- | | colVars | Supports data.frame | | colmeans | Supports data.frame | | invdir.mle | A small optimization |
Improved (by speed, correctness or options)
| Function | What's new! | | ------------------- | ------------------------------- | | Dist(method = "bhattacharyya") | Minor optimization. |
Improved (by speed, correctness or options)
| Function | What's new! | | ------------------- | ------------------------------- | | Dist, total.dist, vecdist | New distance algorithm named "haversine". | | gaussian.nb | 3 times faster than the old version. Add new option for parallel which decreases execution 18 times. | | checkUsage | Fix a bug about "=function". Optimized empty return values and error messages. | | checkAliases, checkTF | Optimized empty return values. | | edist | Fix a bug. | | dirimultinom.mle | Fix a bug. |
Improved (by speed, correctness or options)
| Function | What's new! | | ------------------- | ------------------------------- | | nth | Fix a bug using na_rm. In github is #52. | | colsums,rowsums | Add new option for removing NAs and optimize option indices. |
Improved (by speed, correctness or options)
| Function | What's new! | | ------------------- | ------------------------------- | | Var | Time improvement for removing NAs.| | checkUsage | From now on check for S3methods and fix bug. | | Tcrossprod | Same as R's tcrossprod. | | Crossprod | Same as R's crossprod. | | matrix_multiplication | Add new arguments for perfoming cross or tcross product. |
Improved (by speed, correctness or options)
| Function | What's new! | | ------------------- | ------------------------------- | | Var | Time improvement for removing NAs.| | colVars | Time improvement and depricate argument "suma".| | rowVars | Time improvement and depricate argument "suma".| | rowMedians | Add option for removing NAs if any.| | colMads | Add option for removing NAs if any, add option for choosing method, fix bug and time improvement.| | rowMads | Add option for removing NAs if any, add option for choosing method, fix bug and time improvement.| | mad2 | Deprecated. Use "Mad" instead.| | med | Deprecated. Use "Median" instead.| | colShuffle | Fix a bug.| | rowShuffle | Fix a bug.| | rmdp | Fix a bug.| | checkUsage | From now on check for S3methods. | | AddToNamespace | From now on export S3methods. | | checkAliases | From now on check the S3methods. | | checkNamespace | Don't use it for now. |
Comments
From now on the Rfast can be used in C++ via "LinkingTo" mechanism. The main namespace is "Rfast". Inside "Rfast" you will find two more namespaces, "vector" and "matrix". Namespace "vector" for calling functions using an Rcpp's or RcppArmadillo's vector. Namespace "matrix" for calling functions using an Rcpp's or RcppArmadillo's matrices. The signatures of the functions and the arguments are the same that are exported in R.
For namespace "vector" the functions that are available are
- median(x)
- var(x, std = false, na_rm = false)
- mad(x, method = "median", na_rm = false)
- shuffle(x,engine = Engine(time(0)) // Engine by default is default_random_engine. You can use anyone from C++.
For namespace "matrix" the functions that are available are
- transpose(x)
- matrix_multiplication(x,y)
- colSort(x, descend = false, stable = false, parallel = false)
- rowSort(x, descend = false, stable = false, parallel = false)
- is_symmetric(x)
- colMedian(x, na_rm = false, parallel = false)
- rowMedian(x, na_rm = false, parallel = false)
- colVars(x, std = false, na_rm = false, parallel = false)
- rowVars(x, std = false, na_rm = false, parallel = false)
- colMads(x, method = "median", na_rm = false, parallel = false)
- rowMads(x, method = "median", na_rm = false, parallel = false)
- colShuffle(x,engine = Engine(time(0)) // Engine by default is default_random_engine. You can use anyone from C++.
- rowShuffle(x,engine = Engine(time(0)) // Engine by default is default_random_engine. You can use anyone from C++.
How to use it
- Just add in "LinkingTo" in your NAMESPACE file the "Rfast" or in Rstudio add in the file "//[[Rcppdepends(Rfast)]]".
Include in your cpp files the header "Rfast.h" and enjoy!
Improved (by speed, correctness or options)
| Function | What's new! | | ------------------- | ------------------------------- | | spml.mle | Fix of a bug. | | dirknn | Time improvement. | | glm_logistic | Made the code more robust. | | vmf.mle | Made the code more robust. | | cor.fbed | Fix of a bug. | | Dist | Fix of a bug. |
Improved (by speed, correctness or options)
| Function | What's new! | | ------------------- | ------------------------------- | | allbetas | Time improvement | | cor.fbed | Time improvement | | cor.fsreg | Time improvement | | omp | Time improvement | | ompr | Time improvement | | score.betaregs | Time improvement | | score.gammaregs | Time improvement | | score.glms | Time improvement | | score.invgaussregs | Time improvement | | score.negbinregs | Time improvement | | score.ztpregs | Time improvement | | group.any | Deprecated and replaced by group | | group.all | Deprecated and replaced by group | | group.min | Deprecated and replaced by group | | group.max | Deprecated and replaced by group | | group.min_max | Deprecated and replaced by group | | group.mean | Deprecated and replaced by group | | group.med | Deprecated and replaced by group | | group.mad | Deprecated and replaced by group | | group.var | Deprecated and replaced by group | | group.sum | Deprecated and replaced by group | | groupcolVars | Deprecated and replaced by "Rfast2colGroup(...,method="var")" | | sort_mat | Deprecated and replaced by "colSort" and "rowSort" | | AddToNamespace | remove unused option | | nth | fix bug |
Improved (by speed, correctness or options)
| Function | What's new! | | ------------------- | ----------------------------------------- | | omp | Time improvement. | | sort_mat | Replaced by "colSort" and "rowSort" and will be removed in the next udate. | | checkUsage | Remove uneccessary option. | | read.examples | Remove uneccessary option. | | checkTF | Remove uneccessary option. | | checkAliases | Remove uneccessary option. | | comb_n | Add option for return list or matrix. | | rownth | Fix of a bug. |
New
| Function | What's new! | | ------------------- | ----------------------------------------- | | omp | Multinomial regression now added. | | omp | Option to standardise the predictor variables. | | cor.fbed | Option to standardise the predictor variables. | | cor.fsreg | Option to standardise the predictor variables. | | el.test2 | Empirical likelihood test for two sample means. |
Improved (by speed, correctness or options)
| Function | What's new! | | ------------------- | ----------------------------------------- | | omp | Time improvement. | | ompr | Time improvement. | | cor.fbed | Time improvement. | | multinom.mle | Time improvement. | | weib.reg | Fix a bug and time improvement. | | spatmed.reg | Time improvement. | | invgauss.reg | Time improvement. | | invgauss.regs | Time improvement. | | gammareg | Time improvement. | | gammaregs | Time improvement. | | gammacon | Time improvement. | | colvm.mle | Time improvement. | | gammaregs | Time improvement. | | el.test1 | Time improvement. | | Norm | Fix a bug. | | which_isFactor | Removed and replaced by "which.is". | | checkUsage | Fix a bug. | | colsums | Time improvement for integer matrices. | | rowsums | Time improvement for integer matrices. | | med | Faster version.| | sort.unique.length | slightly faster version.| | sort.unique | slightly faster version.| | Stack | Fix a bug and add function clear for efficient reuse of the Stack. | | read.example | Fix a bug. |
New
| Function | What's new! | | ------------------- | ----------------------------------------- | | omp | Multinomial regression now added. |
Comments
From now on the Rfast needs R version >= 3.5
Improved (by speed, correctness or options)
| Function | What's new! | | ------------------- | -------------------- | | ompr | Time improvement. | | omp | Time improvement . | | cholesky | Time improvement. | | colnth, rownth | Add extra options, "na.rm" and "descending" order and "index.return". | | colrow.zero | deprecate it and replaced by "colrow.value". |
New
| Function | What's new! | | ---------------- | --------------------------------------- | | which.is | The same with which_isFactor but general. Use this instead of which_isFactor. | | colrow.value | Search if a column and row is filled with a specific value. |
Improved (by speed, correctness or options)
| Function | What's new! | | ---------------- | ----------------- | | topological_sort | Time improvement. | | permcor | Fix a bug. | | omp | Time improvement. | | normlog.regs | Fix a bug. | | cova | Added an extra argument. | | ompr | Time improvement. | | as_integer | Fix a bug. | | Round | Fix a bug. digit argument can be up to 15. | | matrix.sum | Deprecate it. | | countNA | Deprecate it. | | sort_unique.length | Deprecate it for numeric numbers. | | Rank | Deprecate method="first". | | Match | Improved. | | nth | Improved method for integers. | | colshuffle | Fix a bug. | | transpose | Now can handle generic matrix using parallel| | colshuffle | Fix a bug. |
New
| Function | What's new! | | ---------------- | ------------------------------------------------------------- | | betabinom.mle | MLE of the beta binomial distribution. | | betageom.mle | MLE of the beta geometric distribution. | | multivt.mle | MLE of the multivariate t distribution. | | colpoisson.anovas | Column-wise ANOVA with Poisson distribution. | | colquasipoisson.anovas | Column-wise ANOVA with quasi Poisson. | | exact.ttest2 | Exact permutations 2-sample t-test. | | chi2Test | Chi-squared test of independence. | | gchi2Test | G-square and Chi-square tests of indepdence. | | chi2tests | Many chi-squared tests of independence. | | chi2Test_univariate | Matrix with chi-square tests of indepedence. | | mvlnorm.mle | MLE of the multivariate lognormal distribution. | | poly.cor | Polychoric correlation. | | pooled.cov | Pooled covariance matrix. | | spatmed.reg | Spatial median (multivariate) regression. | | sscov | Spatial sign covariance matrix. | | trim.colmeans | Trimmed column-wise means. | | trim.rowmeans | Trimmed row-wise means. | | eigs.sym | Extract some principal components from a symmetric matrix. | | invgauss.regs | Many simple inverse Gaussian regressions with a log link. | | invgauss.reg | Inverse Gaussian regression with a log link. | | gammaregs | Many simple Gamma regressions with a log link. | | gammareg | Gamma regression with a log link. | | gammacon | Gamma regression with a constant term only. | | Outer | like R's outer. | | RemoveFromNamespace | remove exported functions. | | Sort.int | fast sorting integer. | | colCumMaxs | apply cummax to column. | | colCumSums | apply cumsum to column. | | colCumMins | apply cummin to column. | | colCumProds | apply cumprod to column. | | positive | apply method to each positive value. | | positive.negative | apply method to each positive and negative value. | | negative | apply method to each negative value. | | as.Rfast.function | convert an R function to Rfast's equivalent | | mat.mult | Generic matrix multiplication using parallel. | | checkUsage | checking usage section in Rd files. | | Hash | Create Hash object. | | Hash.key | Search key or multi key. | | apply.condition | Apply method to each column using a condition. Only integers. | | Stack | Stack object. See man page. | | iterator | iterator object. See man page. | | Elem | access element of an iterator object. | | print.environment | S3 method for printing environment. | | env.copy | deep copy environment. | | ufactor | Untyped factor object. See man page. |
Improved (by speed, correctness or options)
| Function | What's new! | | ---------------- | ----------------- | | rowMedians | Fix a bug | | mad2 | Add option handling NAs and fix a bug | | prop.regs | Made it more stable. | | logistic_only | Made it more stable. | | multinom.regs | Fix a bug and removed parallel for safety reasons. | | rint.regs | Fix a bug. | | univglms | Add the option for quasi Poisson regression | | hd.eigen | Now returns eigen-vectors as well. | | ompr | Fix a bug. | | tobit.mle | Made it faster. | | standardise | Made it faster. | | auc | Made it faster. | | colaucs | Made it faster. | | pc.skel | Made it faster and added more utilities. | | cor.fsreg | Made it faster and fix a bug. | | allttests | Made it faster and fix a bug. | | data.frame.to_matrix | Add option setting colnames and rownames. Fix a bug | | nth | Add option for handling NAs. | | Pmax | Add option for handling NAs. | | Pmin | Add option for handling NAs. | | Sort | Add option for handling NAs. | | Table | Add option for handling NAs, add option for second argument, deprecate argument as.vector and replaced from "names". | | Round | Fix a bug. | | Norm | Fix a bug. | | colsums | Add option for sum using specific indices. | | rowsums | Add option for sum using specific indices and option for parallel. | | is_element | Fix a bug. | | eachrow | Deprecate argument suma and replaced from argument method. | | permutation | Deprecate argument all and replaced from nperm. | | permutation.next | Deprecate argument all.next and replaced from nperm. | | permutation.prev | Deprecate argument all.prev and replaced from nperm. | | data.frame.to_matrix | Fix a bug and time improvement. | | Rank | Fix a bug for method="first". | | Match | fix bug and time improvement. | | bc | Estimate the optimal lambda in the Box-Cox transformation. | | ar1 | AR(1) model. | | colar1 | Many column-wise AR(1) models. | | bc | Estimate the optimal lambda in the Box-Cox transformation. | | rbing | Random values generation from a special case of the Bingham distribution. | | rbingham | Random values generation from the Bingham distribution. | | omp | Orthogonal Matching Pursuit allowing many regression models. | | yule | Yule's coefficient of colligation. | | col.yule | Many column-wise Yule's coefficients of colligation. | | cox.poisrat | Test for the ratio of two Poisson means. | | col.coxpoisrat | Many column-wise tests for the ratio of two Poisson means. | | colPmax | column-wise parallel maxima | | colPmin | column-wise parallel minima | | freq.max | maximum frequency of a number | | freq.min | minimum frequency of a number | | Pmin_Pmax | parallel minima-maxima values | | Table.sign | counting the positive, negative, zeros and NA values. | | topological_sort | Topological sort of a Directed Acyclic Graph (DAG) | | countNA | count the NAs | | columns | get specific columns from a matrix | | rows | get specific rows from a matrix | | eachcol.apply | apply a function to each col after the operation | | checkTF | checking man files for missing TRUE/FALSE values in examples |