Skip to content

In this project, we focus on different ways to optimize a machine learning model parameters.

License

Notifications You must be signed in to change notification settings

pouyaardehkhani/ParameterOptimization

Repository files navigation

ParameterOptimization license releases

What I focus in this project:

  1. Difference between parameters and hyper-parameters. Open In Colab
  2. Different Metrics.
  3. Different ways of Cross Validation.
  4. Basic Search Open In Colab
    • Manual Search
    • Grid Search
    • Random Search
  5. Bayesian Optimisation Open In Colab
    • Sequential Search
    • Trade-off
    • Statistics
    • Bayes’ Rule
    • Probability reallocation
    • Gaussian Distribution
    • Gaussian Process
    • Kernels
    • Acquisition Functions
    • Implementation
  6. Other SMBO Algorithms Open In Colab
    • Bayesian Optimization with Random Forests (SMAC)
    • Tree-structured Parzen Estimators (TPE)
    • Search strategies
    • Annealing with Hyperopt
  7. Scikit-Optimize Open In Colab
    • Search algorithms
    • objective function
    • Sklearn
    • Search Space
    • Acquisition Function
    • Analysis
    • parallelization
    • Implementation
  8. Hyperopt Open In Colab
    • Search algorithms
    • objective function
    • Search Space Configuration
    • nested spaces
    • Acquisition Function
    • Analysis: Trials
    • Parallelization -MongoDB
    • Implementation
  9. Optuna Open In Colab
    • Search algorithms
    • objective function
    • Distributions
    • Acquisition Function
    • Search analysis
    • Parallelization -SQLite
    • Main setup
    • Implementation