项目作者: svecon

项目描述 :
Experiments for comparing Metric Learning algorithms
高级语言: Jupyter Notebook
项目地址: git://github.com/svecon/metric-learn-playground.git
创建时间: 2016-06-16T08:10:49Z
项目社区:https://github.com/svecon/metric-learn-playground

开源协议:

下载


Dimensionality Reduction and Visualization Using Evolutionary Algorithms and Neural Networks

These notebooks contain experiments that are described in my Master thesis “Evolutionary Algorithms for Data Transformation“.

Structure

The experiments are split into several jupyter notebooks:

  • [01]: the datasets used in the paper can be downloaded here, but they are also available in the datasets folder
  • [02]: improving success rate of kNN classification experiment
  • [03]: extracting results and plotting graphs from the [02] experiment
  • [04]: dimensionality reduction experiment
  • [05]: generalization of the evolution experiment
  • [06]: measuring learning times on datasets from [01] experiment
  • [07]: measuring run times on continuously increasing dimensions experiment

The raw results from the experiments are saved in results folder. The resulting graphs are in the graphs folder.