Class BootstrapFitDetectorOptions
Configuration options for the Bootstrap Fit Detector, which evaluates model fit quality using bootstrap resampling.
public class BootstrapFitDetectorOptions- Inheritance
-
BootstrapFitDetectorOptions
- Inherited Members
Remarks
Bootstrap resampling is a statistical technique that creates multiple datasets by randomly sampling with replacement from the original dataset. By training models on these resampled datasets and evaluating their performance, we can assess how well a model generalizes and detect issues like overfitting or underfitting.
For Beginners: This class contains settings for a tool that helps you determine if your AI model is working well. It uses a technique called "bootstrapping" - imagine randomly picking data points from your dataset (sometimes picking the same point multiple times) to create many similar-but-different datasets. By training your model on these different datasets and seeing how consistent the results are, we can tell if your model is learning real patterns or just memorizing the training data.
Properties
ConfidenceInterval
Gets or sets the confidence interval used for statistical assessments.
public double ConfidenceInterval { get; set; }Property Value
- double
The confidence interval as a decimal between 0 and 1, defaulting to 0.95 (95%).
Remarks
The confidence interval represents the probability that the true parameter value falls within the estimated range. A 95% confidence interval (the default) means we are 95% confident that the true value lies within our estimated range.
For Beginners: This setting determines how certain we want to be about our conclusions. The default value of 0.95 means we want to be 95% confident. Think of it like a weather forecast - saying there's a 95% chance of rain means you should probably bring an umbrella. Higher values (like 0.99) mean we want to be even more certain before making a conclusion, while lower values (like 0.90) mean we're willing to be less certain.
GoodFitThreshold
Gets or sets the threshold for identifying a good model fit.
public double GoodFitThreshold { get; set; }Property Value
- double
The good fit threshold as a decimal between 0 and 1, defaulting to 0.9 (90%).
Remarks
If the model's performance exceeds this threshold and there is no significant difference between training and testing performance, the model is considered to have a good fit.
For Beginners: This setting defines what we consider "good performance" for your model. The default value (0.9) means we'll consider your model to be performing well if it achieves at least 90% accuracy and doesn't show signs of overfitting. Think of it as setting the bar for what counts as success. You might lower this if your problem is particularly difficult, or raise it if you need extremely high accuracy.
NumberOfBootstraps
Gets or sets the number of bootstrap samples to generate for evaluation.
public int NumberOfBootstraps { get; set; }Property Value
- int
The number of bootstrap samples, defaulting to 1000.
Remarks
This parameter determines how many resampled datasets will be created and evaluated. A higher number provides more reliable estimates but increases computation time.
For Beginners: This setting controls how many different random samples we'll create to test your model. The default (1000) gives reliable results for most cases. Think of it like flipping a coin - if you flip it only 10 times, you might get 7 heads and think the coin is biased. But if you flip it 1000 times, you'll get a much more accurate picture of whether the coin is fair. More samples give more reliable results but take longer to process.
OverfitThreshold
Gets or sets the threshold for detecting overfitting.
public double OverfitThreshold { get; set; }Property Value
- double
The overfit threshold as a decimal between 0 and 1, defaulting to 0.1 (10%).
Remarks
If the difference between training performance and testing performance exceeds this threshold, the model is considered to be overfitting. Lower values make the detector more sensitive to overfitting.
For Beginners: Overfitting happens when your model performs great on training data but poorly on new data - it's like memorizing exam questions instead of understanding the subject. This setting determines how big the difference between training and testing performance needs to be before we warn you about overfitting. The default value (0.1) means if your model is more than 10% better on training data than on test data, we'll flag it as potentially overfitting.
UnderfitThreshold
Gets or sets the threshold for detecting underfitting.
public double UnderfitThreshold { get; set; }Property Value
- double
The underfit threshold as a decimal between 0 and 1, defaulting to 0.7 (70%).
Remarks
If the model's performance (on both training and testing data) is below this threshold, the model is considered to be underfitting. Higher values make the detector more sensitive to underfitting.
For Beginners: Underfitting happens when your model is too simple to capture the patterns in your data - like trying to draw a circle using only straight lines. This setting determines the minimum performance level your model needs to achieve to avoid being flagged as underfitting. The default value (0.7) means if your model's accuracy is below 70% on both training and test data, we'll suggest that it might be underfitting and needs to be more complex.