NNS-Python

Nonlinear Nonparametric Statistics

Install

pip install NNS

Implemented Functions:

From Beta R Version of 2021-12-13 (Version: 8.4-Beta, Date: 2021-12-13)

• ANOVA

  • NNS.ANOVA: TODO (deps: NNS.ANOVA.bin)

• ARMA

  • NNS.ARMA: TODO (deps: NNS.seas, ARMA.seas.weighting, NNS.meboot)

• ARMA_optim:

  • NNS.ARMA.optim: TODO (deps: NNS.ARMA)

• Binary_ANOVA

  • NNS.ANOVA.bin: OK

• Boost

  • NNS.boost: TODO (deps: NNS.caus, NNS.reg, NNS.stack)

• Causal_Matrix

  • NNS.caus.matrix: TODO (deps: NNS.caus)

• Causation

  • NNS.caus: TODO (deps: Uni.caus, NNS.caus.matrix)

• Copula

  • NNS.copula: OK

• Dependence

  • NNS.dep: TODO (deps: NNS.part, NNS.dep.matrix)

• Dependence_matrix

  • NNS.dep.matrix: TODO (deps: NNS.dep)

• dy_d_wrt

  • dy.d_: TODO (deps: NNS.reg)

• dy_dx

  • dy_dx: TODO (deps: NNS.dep, NNS.reg)

• Internal Functions

  • mode: TEST
  • mode_class: TODO
  • gravity: TEST
  • gravity_class: TODO
  • factor_2_dummy: TODO
  • factor_2_dummy_FR: TODO
  • generate_vectors: TODO
  • ARMA_seas_weighting: TODO
  • is.discrete: TODO
  • lag_mtx: TODO
  • NNS_meboot_part: TODO
  • NNS_meboot_expand_sd: TODO
  • alt_cbind: TEST (not in newest version, maybe R related)
  • RP: TODO (not in newest version)

• LPM UPM VaR

  • LPM_VaR: OK
  • UPM_VaR: OK
  • used np.quantile instead of tdigest, and root_scalar instead of optimize

• Multivariate_Regression

  • NNS.M.reg: TODO (deps: NNS.part, NNS.dep, NNS::NNS.distance, NNS.copula, NNS.reg)

• NNS_Distance

  • NNS.distance: TODO (deps: dtw, Rfast)

• NNS_meboot

  • NNS.meboot: TODO (deps: NNS.dep, NNS.meboot.expand.sd)

• NNS_term_matrix

  • NNS.term.matrix: OK

• NNS_VAR

  • NNS.VAR: TODO (deps: NNS.reg, NNS.seas, NNS.ARMA.optim, NNS.ARMA, NNS.stack, NNS.dep, NNS.caus)

• Normalization

  • NNS.norm: TODO (deps: NNS.dep, Rfast)

• Nowcast

  • NNS.nowcast: TODO (deps: Quandl, NNS.VAR)

• Numerical Differentiation

  • NNS.diff: TODO (nodeps)

• Partition_Map

  • NNS.part: TODO (deps: internal functions: gravity_class, gravity, mode_class)

• Partial Moments

  • pd_fill_diagonal: OK (Internal use)
  • LPM: OK Tested
    • numba_LPM: Numba version (Internal use)
    • LPM: Vectorized / pandas / numpy friendly
  • UPM: OK Tested
    • numba_UPM: Numba version (Internal use)
    • UPM: Vectorized / pandas / numpy friendly
  • Co_UPM: OK Tested
    • _Co_UPM: Internal Use
    • _vec_Co_UPM: numpy.vectorized
    • Co_UPM: Vectorized / pandas / numpy friendly
  • Co_LPM: OK Tested
    • _Co_LPM: Internal Use
    • _vec_Co_LPM: numpy.vectorized
    • Co_LPM: Vectorized / pandas / numpy friendly
  • D_LPM: OK Tested
    • _D_LPM: Internal User
    • _vec_D_LPM: numpy.vectorized
    • D_LPM: Vectorized / pandas / numpy friendly
  • D_UPM: OK Tested
    • _D_UPM: Internal User
    • _vec_D_UPM: numpy.vectorized
    • D_UPM: Vectorized / pandas / numpy friendly
  • PM_matrix: OK
  • LPM_ratio: OK
  • UPM_ratio: OK
  • NNS_PDF: TODO (deps: d/dx approximation, density)
  • NNS_CDF: TODO (deps: ecdf, density, matplotlib, NNS_reg)

• Regression

  • NNS.reg: TODO (deps: NNS.M.reg, NNS.dep, NNS.part, Uni.caus)

• SD Efficient Set

  • NNS_SD_efficient_set: OK (TODO: numba version?)

• Seasonality_Test

  • NNS.seas: TODO (nodeps)

• Stack

  • NNS.stack: TODO (deps: NNS.reg, NNS::NNS.distance)

• Uni_Causation

  • Uni.caus: TODO (deps: NNS.norm, NNS.dep)

• FSD, SSD, TSD

  • NNS_FSD: OK (TODO: numba version?)
  • NNS_SSD: OK (TODO: numba version?)
  • NNS_TSD: OK (TODO: numba version?)

• Uni SD Routines

  • NNS_FSD_uni: OK (TODO: numba version?)
  • NNS_SSD_uni: OK (TODO: numba version?)
  • NNS_TSD_uni: OK (TODO: numba version?)

• Others Todos:

  • Try to make names equal to R version
    • R accept $ and . we will replace to underline _
    • TODO: R export functions from modules