Trjtdtk

By means of an example, show that P(A) + P(B) = 1 does not mean that B is the complement of A.

Closed-form expression for certain product

Biological Blimps: Propulsion

When were female captains banned from Starfleet?

How to implement a feedback to keep the DC gain at zero for this conceptual passive filter?

Should I stop contributing to retirement accounts?

React - map array to child component

Writing bit difficult equation in latex

250 Floor Tower

Is "staff" singular or plural?

Can I sign legal documents with a smiley face?

Is there a single word describing earning money through any means?

Calculating Wattage for Resistor in High Frequency Application?

Create all possible words using a set or letters

What is this called? Old film camera viewer?

Should I outline or discovery write my stories?

Why should universal income be universal?

What prevents the use of a multi-segment ILS for non-straight approaches?

Is this toilet slogan correct usage of the English language?

Could the E-bike drivetrain wear down till needing replacement after 400 km?

why `nmap 192.168.1.97` returns less services than `nmap 127.0.0.1`?

GraphicsGrid with a Label for each Column and Row

Is there a working SACD iso player for Ubuntu?

What was the exact wording from Ivanhoe of this advice on how to free yourself from slavery?

Hyperparameter Tuning with Simulated Data

Are Support Vector Machines still considered “state of the art” in their niche?Linear kernel in SVM performing much worse than RBF or PolyHow to store and analyze classification results with Python?Automated tuning of HyperparameterHow to deal with unbalanced data in pixelwise classification?Hyperparameter Optimization for a Machine Learning AlgorithmTuning svm and cart hyperparametersHyperparameter tuning for stacked modelsDisadvantages of hyperparameter tuning on a random sample of datasetDocument parsing modeling and approach?

1

$begingroup$

I'm trying to create a SVM classifier which can predict some fault, and to train it I'm using simulated examples of the fault. Of course, the simulations are not perfect, but they appear to be good enough since I get reasonable results when predicting on the real examples.

What is the best way to tune the hyperparameters in order to maximise the performance on the real data? Using GridSearchCV does a good job of maximising the accuracy on a subset of the simulated data, but that doesn't always mean good results on the real data.

Any general tips on working with simulated training data are also welcome.

classification svm hyperparameter-tuning

share|improve this question

asked Mar 19 at 12:03

CanofDrinkCanofDrink

61

New contributor

CanofDrink is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.

$endgroup$

add a comment | 

1

$begingroup$

I'm trying to create a SVM classifier which can predict some fault, and to train it I'm using simulated examples of the fault. Of course, the simulations are not perfect, but they appear to be good enough since I get reasonable results when predicting on the real examples.

What is the best way to tune the hyperparameters in order to maximise the performance on the real data? Using GridSearchCV does a good job of maximising the accuracy on a subset of the simulated data, but that doesn't always mean good results on the real data.

Any general tips on working with simulated training data are also welcome.

classification svm hyperparameter-tuning

share|improve this question

asked Mar 19 at 12:03

CanofDrinkCanofDrink

61

New contributor

CanofDrink is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.

$endgroup$

add a comment | 

$begingroup$

I'm trying to create a SVM classifier which can predict some fault, and to train it I'm using simulated examples of the fault. Of course, the simulations are not perfect, but they appear to be good enough since I get reasonable results when predicting on the real examples.

What is the best way to tune the hyperparameters in order to maximise the performance on the real data? Using GridSearchCV does a good job of maximising the accuracy on a subset of the simulated data, but that doesn't always mean good results on the real data.

Any general tips on working with simulated training data are also welcome.

classification svm hyperparameter-tuning

share|improve this question

asked Mar 19 at 12:03

CanofDrinkCanofDrink

61

New contributor

CanofDrink is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.

$endgroup$

share|improve this question

asked Mar 19 at 12:03

CanofDrinkCanofDrink

61

New contributor

CanofDrink is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.

share|improve this question

asked Mar 19 at 12:03

CanofDrinkCanofDrink

61

New contributor

CanofDrink is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.

asked Mar 19 at 12:03

CanofDrinkCanofDrink

61

New contributor

CanofDrink is a new contributor to this site. Take care in asking for clarification, commenting, and answering.
Check out our Code of Conduct.

asked Mar 19 at 12:03

CanofDrinkCanofDrink

61