sgGWR package

Subpackages

• sgGWR.optimizers package
  • Submodules
  • sgGWR.optimizers.existings module
    • optax_optimizer
      • optax_optimizer.run()
    • scipy_L_BFGS_B
      • scipy_L_BFGS_B.run()
  • sgGWR.optimizers.existings_numpy module
    • scipy_L_BFGS_B
      • scipy_L_BFGS_B.run()
  • sgGWR.optimizers.golden module
    • golden_section
      • golden_section.run()
  • sgGWR.optimizers.second module
    • SGN
      • SGN.run()
      • SGN.step()
    • SGN_BFGS
      • SGN_BFGS.step()
    • SGN_LM
      • SGN_LM.step()
  • sgGWR.optimizers.sg module
    • ASGD
      • ASGD.lr_schedule()
      • ASGD.step()
    • Adam
      • Adam.step()
    • SGD
      • SGD.lr_schedule()
      • SGD.run()
      • SGD.step()
    • SGDarmijo
      • SGDarmijo.armijo_cond()
      • SGDarmijo.armijo_search()
      • SGDarmijo.lr_schedule()
      • SGDarmijo.step()
    • Yogi
      • Yogi.step()
  • sgGWR.optimizers.sg_numpy module
    • ASGD
      • ASGD.lr_schedule()
      • ASGD.step()
    • SGD
      • SGD.lr_schedule()
      • SGD.run()
      • SGD.step()
    • SGDarmijo
      • SGDarmijo.armijo_cond()
      • SGDarmijo.armijo_search()
      • SGDarmijo.lr_schedule()
      • SGDarmijo.step()
  • sgGWR.optimizers.vr module
    • KatyushaXs
    • KatyushaXw
    • SVRG
      • SVRG.batch_step()
      • SVRG.run()
      • SVRG.step()
  • Module contents
    • ASGD
      • ASGD.lr_schedule()
      • ASGD.step()
    • Adam
      • Adam.step()
    • KatyushaXs
    • KatyushaXw
    • SGD
      • SGD.lr_schedule()
      • SGD.run()
      • SGD.step()
    • SGDarmijo
      • SGDarmijo.armijo_cond()
      • SGDarmijo.armijo_search()
      • SGDarmijo.lr_schedule()
      • SGDarmijo.step()
    • SGN
      • SGN.run()
      • SGN.step()
    • SGN_BFGS
      • SGN_BFGS.step()
    • SGN_LM
      • SGN_LM.step()
    • SVRG
      • SVRG.batch_step()
      • SVRG.run()
      • SVRG.step()
    • Yogi
      • Yogi.step()
    • golden_section
      • golden_section.run()
    • optax_optimizer
      • optax_optimizer.run()
    • scipy_L_BFGS_B
      • scipy_L_BFGS_B.run()

Submodules

sgGWR.kernels module

class sgGWR.kernels.Biweight(params)

Bases: _scaledKernel

class sgGWR.kernels.Epanechnikov(params)

Bases: _scaledKernel

class sgGWR.kernels.Exponential(params)

Bases: _scaledKernel

class sgGWR.kernels.Gaussian(params)

Bases: _scaledKernel

class sgGWR.kernels.LinearMultiscale(sites, params=Array([1.e-04, 1.e+00], dtype=float32), base_kernel=None, n_poly=4, n_neighbour=100)

Bases: _KDTreeKernel

dk(x1, x2, params)

k(x1, x2, params)

class sgGWR.kernels.Triangular(params)

Bases: _scaledKernel

class sgGWR.kernels.stBiweight(params)

Bases: _scaledSTKernel

class sgGWR.kernels.stEpanechnikov(params)

Bases: _scaledSTKernel

class sgGWR.kernels.stExponential(params)

Bases: _scaledSTKernel

class sgGWR.kernels.stGaussian(params)

Bases: _scaledSTKernel

class sgGWR.kernels.stTriangular(params)

Bases: _scaledSTKernel

sgGWR.models module

class sgGWR.models.GWR(y, X, sites, kernel=<sgGWR.kernels.Gaussian object>)

Bases: GWR_Ridge

grad_params_aicc(params, idx=None, sigma2_type=0)

grad_params_loocv(params, idx=None)

loocv_GN(params, idx=None)

loocv_loss(params, idx=None)

set_params(unconstrained, transform=True)

unconstrained_GN(x, idx=None)

unconstrained_grad(x, idx=None)

unconstrained_loss(x, idx=None)

class sgGWR.models.GWR_Ridge(y, X, sites, kernel=<sgGWR.kernels.Gaussian object>, penalty=0.01)

Bases: object

AICc(params=None, sigma2_type=0)

Fast Evaluation of AICc

reference. Li, Z., Fotheringham, A. S., Li, W., & Oshan, T. (2019). Fast Geographically Weighted Regression (FastGWR): a scalable algorithm to investigate spatial process heterogeneity in millions of observations. International Journal of Geographical Information Science, 33(1), 155–175.

get_beta(s)

grad_params_aicc(params, penalty, idx=None, sigma2_type=0)

grad_params_loocv(params, penalty, idx=None)

grad_penalty_aicc(params, penalty, idx=None)

grad_penalty_loocv(params, penalty, idx=None)

loocv_GN(params, penalty, idx=None)

loocv_loss(params, penality, idx=None)

setInferenceStats(alpha=0.05)

set_betas_inner()

set_params(unconstrained, transform=True)

unconstrained_GN(x, idx=None)

unconstrained_grad(x, idx=None)

unconstrained_loss(x, idx=None)

class sgGWR.models.MGWR(y, X, sites, kernel=<sgGWR.kernels.Gaussian object>, base_class=<class 'sgGWR.models.GWR_Ridge'>, base_class_params={})

Bases: GWR_Ridge

backfitting(optimizers, maxiter=100, verbose=True, tol=1e-05, run_params={})

class sgGWR.models.ScaGWR(y, X, sites, kernel, precompute=True)

Bases: GWR

grad_params_loocv(params, idx=None)

Module contents