PySensors examples

Here we provide examples of how to use PySensors objects to solve sensor placement problems.

PySensors overview

This notebook gives an overview of most of the different tools available in PySensors. It’s a good place to start to get a quick idea of what the package is capable of.

Basis comparison

This example compares the different basis options implemented in PySensors on a simple problem.

Classification with SSPOC

See how to use the SSPOC class (Sparse Sensor Placement Optimization for Classification) to choose sparse sets of sensors for classification problems.

Cost constraints

Learn about the CCQR optimizer and how it can be used to place sparse sensors when there are variable costs associated with different locations.

Cross validation

PySensors was designed to be completely compatible with scikit-learn. In this notebook we show how to perform cross-validation with scikit-learn objects to optimize the number of sensors and/or basis modes.

Sea surface temperature prediciton

See how PySensors can be used to pick optimal locations for sensors in the ocean to help predict the temperature of the ocean at any given point.

Vandermonde example

Reproduces an example from Manohar et al. (2018) where sensor locations are learned for a monomial basis for the task of reconstruction.

Full table of contents

• Cost-constrained QR (CCQR)
  • Setup
  • Sensor costs
• Sea Surface Temperature (SST) sensors
  • Load data
    • Plot a snapshot
  • Optimal sensor placement
• Basis comparison
  • Setup
  • Reconstruction error
    • Varying the number of basis modes
    • Varying the number of sensors
  • Sensor locations
• PySensors Overview
  • Reconstruction
    • Setup
    • Simplest case
    • Changing the number of sensors
    • Forming reconstructions
    • Sweeping over a range of sensor counts
    • Changing the basis
      • SVD basis
      • Random projections
      • Updating the number of basis modes
    • Cost-constraints
  • Classification
    • Simplest case
    • Changing the number of sensors
    • Changing the basis
    • Changing the classifier
    • Measuring accuracy
• Cross-validation
  • Setup
  • Grid search
  • Randomized search
• Sparse Sensor Placement Optimization for Classification (SSPOC)
  • Binary classification
  • Multiclass classification
    • Choosing sensors
      • Top-k sensors
      • Thresholding
    • Varying the L1 penalty
  • Other classifiers
  • Other bases
    • Varying the number of sensors
    • Varying the number of basis modes
• Spatial Constraints
• Setup
• Unconstrained optimization of sensor placement:
• The exact_n case:
• Reconstruct image from test set using sensors placed via constrained (exact_n) optimizer
  • Compare reconstruction errors on test set for unconstrained vs. constrained exact_n sensor placement:
  • Observations:
  • Conclusions:
• The max_n case:
  • Conclusions:
• The predetermined case:
• Reconstruct image from test set using sensors placed via predetermined optimizer
  • Compare the reconstruction errors for unconstrained and predetermined sensor placement on the test set:
• Polynomial interpolation

Cost-constrained QR (CCQR)

Sea Surface Temperature (SST) sensors

Basis comparison

PySensors Overview

Cross-validation

Sparse Sensor Placement Optimization for Classification (SSPOC)

Spatial Constraints

Polynomial interpolation