Comparison of Gaussian process software: Difference between revisions

Revision as of 21:29, 2 March 2025

Revision as of 22:10, 2 March 2025

Line 548:

| [[Python (programming language)|Python]]

| {{yes}}

| {{Yesyes|Heteroskedastic, VAE, POD}}{{efn|name=POD| POD (Proper Orthogonal Decomposition) is a dimensionality reduction technique used in Gaussian Process regression to approximate complex systems by projecting data onto a lower-dimensional subspace, making computations more efficient. It assumes the system is governed by a few dominant modes, making it ideal for problems with clear separability of scales, but less effective when all dimensions contribute equally to the system's behavior.<ref name="Porrello24" />}}

| {{yes|POD}}{{efn|name=POD}}

| {{yes|Sparse}}

| {{yes|ND}}

Line 594:

| {{yes}}

! [https://github.com/wesselb/stheno Stheno]

|-

![https://CODES.arizona.edu CODES]

|

|[[MATLAB]]

|{{Yes}}

|{{Yes|Heteroskedastic, VAE, POD}}{{efn|name=POD| POD (Proper Orthogonal Decomposition) is a dimensionality reduction technique used in Gaussian Process regression to approximate complex systems by projecting data onto a lower-dimensional subspace, making computations more efficient. It assumes the system is governed by a few dominant modes, making it ideal for problems with clear separability of scales, but less effective when all dimensions contribute equally to the system's behavior.<ref name="Porrello24" />}}

|{{Yes|Sparse}}

|{{Yes|ND}}

|{{No}}

|{{No|Gaussian}}

|{{partial|i.i.d}}

|{{partial|Some, Automatic}}

|{{partial|Mean Aposteriori}}

|{{No}}

|{{No}}

|{{No}}

![https://CODES.arizona.edu CODES]

|-

! [https://docs.rs/egobox-gp/latest/egobox_gp/ Egobox-gp]<ref name="Lafage2022" />