Class AutocorrelationFitDetectorOptions

Namespace

AiDotNet.Models.Options

Assembly

AiDotNet.dll

Configuration options for detecting autocorrelation in time series data and regression residuals.

public class AutocorrelationFitDetectorOptions

Inheritance

object

AutocorrelationFitDetectorOptions

Inherited Members

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.MemberwiseClone()

object.ReferenceEquals(object, object)

object.ToString()

Remarks

This class provides threshold values used to interpret the Durbin-Watson statistic, which measures autocorrelation in the residuals (errors) of regression and time series models. The Durbin-Watson statistic typically ranges from 0 to 4, with different values indicating different types of autocorrelation.

For Beginners: Autocorrelation is a pattern where data points in a time series are related to their own past values. Think of it like weather patterns - if it's been raining for several days, there's a higher chance it will rain tomorrow (positive autocorrelation). This class helps determine if your data has such patterns by setting thresholds for the Durbin-Watson test, which is like a thermometer for measuring autocorrelation. Values around 2 suggest no autocorrelation, values closer to 0 suggest positive autocorrelation (each value tends to be similar to previous values), and values closer to 4 suggest negative autocorrelation (each value tends to be opposite to previous values). Understanding autocorrelation helps choose the right prediction model for your data.

Properties

NoAutocorrelationLowerBound

Gets or sets the lower bound of the range indicating no autocorrelation.

public double NoAutocorrelationLowerBound { get; set; }

Property Value

double

The lower bound value, defaulting to 1.5.

Remarks

Durbin-Watson statistic values above this threshold and below the upper bound suggest that there is no significant autocorrelation in the data.

For Beginners: This is the minimum value for the "safe zone" where your data likely doesn't have significant autocorrelation patterns. If your test result is between this lower bound (1.5 by default) and the upper bound, it suggests your data points aren't strongly influenced by previous values. This is often ideal for many standard statistical methods, as they typically assume independence between observations.

NoAutocorrelationUpperBound

Gets or sets the upper bound of the range indicating no autocorrelation.

public double NoAutocorrelationUpperBound { get; set; }

Property Value

double

The upper bound value, defaulting to 2.5.

Remarks

Durbin-Watson statistic values below this threshold and above the lower bound suggest that there is no significant autocorrelation in the data.

For Beginners: This is the maximum value for the "safe zone" where your data likely doesn't have significant autocorrelation patterns. If your test result is between the lower bound and this upper bound (2.5 by default), it suggests your data points aren't strongly influenced by previous values. When your test result falls in this range (approximately 1.5-2.5), you can generally proceed with standard statistical methods without special adjustments for autocorrelation.

StrongNegativeAutocorrelationThreshold

Gets or sets the threshold for detecting strong negative autocorrelation.

public double StrongNegativeAutocorrelationThreshold { get; set; }

Property Value

double

The threshold value, defaulting to 3.0.

Remarks

Durbin-Watson statistic values above this threshold indicate strong negative autocorrelation in the data. Negative autocorrelation means that values tend to be followed by opposite values.

For Beginners: This is a threshold for detecting patterns where each data point tends to be opposite to the previous one. If the test result is above 3.0 (the default threshold), it suggests your data has strong negative autocorrelation - meaning high values tend to be followed by low values and vice versa (like a very hot day often being followed by a cooler day). This zigzag pattern requires special consideration when building prediction models.

StrongPositiveAutocorrelationThreshold

Gets or sets the threshold for detecting strong positive autocorrelation.

public double StrongPositiveAutocorrelationThreshold { get; set; }

Property Value

double

The threshold value, defaulting to 1.0.

Remarks

Durbin-Watson statistic values below this threshold indicate strong positive autocorrelation in the data. Positive autocorrelation means that values tend to be followed by similar values.

For Beginners: This is like a warning level for detecting patterns where each data point is similar to the previous one. If the test result is below 1.0 (the default threshold), it suggests your data has strong positive autocorrelation - meaning values tend to follow similar values (like warm days typically following warm days). This pattern affects how you should analyze and predict your data, as simple models might not capture this relationship correctly.