EDDM: EDDM (Early Drift Detection Method)

In datadriftR: Concept Drift Detection Methods for Stream Data

EDDM (Early Drift Detection Method)

Description

This class implements the Early Drift Detection Method (EDDM), designed to detect concept drifts in online learning scenarios by monitoring the distances between consecutive errors. EDDM is particularly useful for detecting gradual drifts earlier than abrupt changes.

Details

EDDM is a statistical process control method that is more sensitive to changes that happen more slowly and can provide early warnings of deterioration before the error rate increases significantly.

Public fields

eddm_warning

Warning threshold setting.

eddm_outcontrol

Out-of-control threshold setting.

m_num_errors

Current number of errors encountered.

m_min_num_errors

Minimum number of errors to initialize drift detection.

m_n

Total instances processed.

m_d

Distance to the last error from the current instance.

m_lastd

Distance to the previous error from the last error.

m_mean

Mean of the distances between errors.

m_std_temp

Temporary standard deviation accumulator for the distances.

m_m2s_max

Maximum mean plus two standard deviations observed.

delay

Delay count since the last detected change.

estimation

Current estimated mean distance between errors.

warning_detected

Boolean indicating if a warning has been detected.

change_detected

Boolean indicating if a change has been detected.

Methods

Public methods

• EDDM$new()

• EDDM$reset()

• EDDM$add_element()

• EDDM$clone()

---

Method new()

Initializes the EDDM detector with specific parameters.

Usage

EDDM$new(min_num_instances = 30, eddm_warning = 0.95, eddm_outcontrol = 0.9)

Arguments

min_num_instances

Minimum number of errors before drift detection starts.

eddm_warning

Threshold for warning level.

eddm_outcontrol

Threshold for out-of-control level.

---

Method reset()

Resets the internal state of the EDDM detector.

Usage

EDDM$reset()

---

Method add_element()

Adds a new observation and updates the drift detection status.

Usage

EDDM$add_element(prediction)

Arguments

prediction

Numeric value representing a new error (usually 0 or 1).

---

Method clone()

The objects of this class are cloneable with this method.

Usage

EDDM$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

References

Early Drift Detection Method. Manuel Baena-Garcia, Jose Del Campo-Avila, Raúl Fidalgo, Albert Bifet, Ricard Gavalda, Rafael Morales-Bueno. In Fourth International Workshop on Knowledge Discovery from Data Streams, 2006.

Implementation: https://github.com/scikit-multiflow/scikit-multiflow/blob/a7e316d1cc79988a6df40da35312e00f6c4eabb2/src/skmultiflow/drift_detection/eddm.py

Examples

set.seed(123)  # Setting a seed for reproducibility
data_part1 <- sample(c(0, 1), size = 100, replace = TRUE, prob = c(0.7, 0.3))

# Introduce a change in data distribution
data_part2 <- sample(c(0, 1), size = 100, replace = TRUE, prob = c(0.3, 0.7))

# Combine the two parts
data_stream <- c(data_part1, data_part2)
eddm <- EDDM$new()
for (i in 1:length(data_stream)) {
  eddm$add_element(data_stream[i])
  if (eddm$change_detected) {
    message(paste("Drift detected!",i))
  } else if (eddm$warning_detected) {
    message(paste("Warning detected!",i))
  }
}

datadriftR documentation built on April 3, 2025, 6:13 p.m.