EQ Feature Engineering

The EQ Feature Engineering node generates earthquake-related attributes from a dataset of seismic events. It uses different approaches to compute seismic indicators that can be used in machine learning models for earthquake prediction.

This node takes a table of seismic events as input and generates seismic attributes based on configurable parameters.

Input Parameters

1. Events for b-value Morales (nMorales):

  • Number of previous events considered for calculating the b-value using Morales’ method.

  • Recommended values: 10 - 100 (default: 10).

2. Events for b-value Adeli (nAdeli):

  • Number of previous events considered for calculating the b-value using Adeli’s method.

  • Recommended values: 1 - 10 (default: 1).

3. Reference Magnitude (referenceMagnitude):

  • Reference magnitude used in b-value calculation.

  • Recommended value: 3.0.

4. Days for Prediction (dayspred):

  • Number of future days considered for prediction.

  • Recommended value: 5 - 7 (default: 6).

5. Classification: From (classFrom):

  • Minimum magnitude for discrete class creation.

  • Recommended value: 2.0 - 5.0.

6. Classification: To (classTo):

  • Maximum magnitude for discrete class creation.

  • Recommended value: 2.0 - 6.0.

7. Classification Step (classStep):

  • Increment between magnitude classes.

  • Recommended values: 0.05 - 0.5.

8. Threshold for mu and c (chth):

  • Characteristic threshold used in the calculation of mu and c attributes.

  • Recommended values: 0.01 - 1.0.

9. Attribute Sets:

  • Defines the set of generated attributes.

  • Options:

    • attYorch/bM - Yorch attributes with Morales’ b-value.
    • attYorch/bA - Yorch attributes with Adeli’ b-value.
    • attAdeli/bM - Adeli attributes with Morales’ b-value.
    • attAdeli/bA - Adeli attributes with Adeli’s b-value.

Output Attributes

The generated attributes include seismic indicators based on b-value calculations and other factors:

  • b-value (bM, bA): Based on Morales’ (bM) or Adeli’s (bA) method.

  • a-value: Gutenberg-Richter law value.

  • b-value increments (x1 - x5): Variation of b-value in previous events.

  • Weekly maximum magnitude (x6): Highest magnitude recorded in the last week.

  • Probability of magnitude >=6.0 (x7): Calculated using probability density distribution.

  • Elapsed time (T): Time since the last significant seismic event.

  • Coefficient of variation (c): Ratio between standard deviation and mean time between events.

  • Mean square deviation (η): Variability in seismic activity.

  • Magnitude deficit (∆M): Difference between observed and expected magnitudes.

  • dE12: This parameter represents the rate of the square root of seismic energy, measuring the speed at which seismic energy is released after applying a square root transformation to reduce its dispersion.

  • Mmean: This is the mean magnitude, calculated as the average of the magnitudes of the events recorded in the catalog.

  • μ: This parameter denotes the mean time, representing the average time interval between seismic events.

Errors and Messages

  • “No data input provided”: A data source must be connected.

  • “Missing columns in input data”: The dataset must contain time, latitude, longitude, and mag.