This is just an experiment arranged by the teacher, the preparation process encountered a variety of problems, now posted for your reference, is Andrew Ng involved in the study,
The article is based on the following, the second is a related paper,
Learning Feature Representations with K-means, Adam Coates and Andrew Y. Ng. In neural networks:tricks of the trade, Reloaded, Springer LNCS, 2012
An analysis of Single-layer Networks in unsupervised Feature learning, Adam Coates, Honglak Lee, and Andrew Y. Ng. In Aistats 14, 2011
The related datasets and programs are run, and are given in the Readme in the file,
Here's the code.
Several related blogs are also given, which correspond to the above papers:
K-means Feature Learning
Analysis of single-layer unsupervised learning Network
Kmeans single-layer network recognition performance gives a clearer explanation,
Using K-means clustering classification to carry out feature learning