Simplified Machine-Learning Workflow #10
Author
Anton Antonov
Title
Simplified Machine-Learning Workflow #10
Description
Semantic Analysis (Part 5)
Category
Educational Materials
Keywords
URL
http://www.notebookarchive.org/2020-09-55swtlo/
DOI
https://notebookarchive.org/2020-09-55swtlo
Date Added
2020-09-11
Date Last Modified
2020-09-11
File Size
14.41 megabytes
Supplements
Rights
Redistribution rights reserved
Download
Open in Wolfram Cloud
Latent Semantic Analysis (Part 5)
Latent Semantic Analysis (Part 5)
A Wolfram livecoding session
Anton Antonov
February 2020
February 2020
Session overview
Session overview
1
.Review: last session’s image collection topics extraction.
1
.1
.Let us try that two other datasets:
1
.1
.1
.handwritten digits, and
1
.1
.2
.Hentaigana (maybe).
2
.Image denoising (maybe):
2
.1
.Using handwritten digits (again).
3
.Image classification:
3
.1
.Handwritten digits.
Data
Data
Course data
Course data
Out[]=
|
In[]:=
WebImage["https://github.com/antononcube/SimplifiedMachineLearningWorkflows-book/tree/master/Data"]
Out[]=
Load packages
Load packages
Review: random mandala images topics
Review: random mandala images topics
LSA over images of handwritten Hentaigana characters (maybe)
LSA over images of handwritten Hentaigana characters (maybe)
LSA over images of handwritten digits
LSA over images of handwritten digits
Get handwritten images data
Get handwritten images data
In[]:=
AbsoluteTiming[lsTrainingData=ExampleData[{"MachineLearning","MNIST"},"TrainingData"];]
Out[]=
{40.7365,Null}
In[]:=
Length[lsTrainingData]
Out[]=
60000
In[]:=
RandomSample[lsTrainingData,12]
Out[]=
6,
1,
6,
9,
1,
7,
8,
3,
4,
9,
5,
1
In[]:=
Tally[lsTrainingData〚All,2〛]
Out[]=
{{0,5923},{1,6742},{2,5958},{3,6131},{4,5842},{5,5421},{6,5918},{7,6265},{8,5851},{9,5949}}
In[]:=
AbsoluteTiming[lsTestData=ExampleData[{"MachineLearning","MNIST"},"TestData"];]
Out[]=
{7.00075,Null}
In[]:=
Length[lsTestData]
Out[]=
10000
In[]:=
SeedRandom[3242];lsTrainingDataSmall=RandomSample[lsTrainingData,4000];lsTestDataSmall=RandomSample[lsTestData,1000];
LSAMon over 5’s only
LSAMon over 5’s only
De-noising: LSA over handwritten images
De-noising: LSA over handwritten images
De-noising: Further comparison with Classify
De-noising: Further comparison with Classify
Classification: LSA classifier for images of handwritten digits
Classification: LSA classifier for images of handwritten digits
Classification: Analogies with Neural networks computations
Classification: Analogies with Neural networks computations
Classification: Further analysis
Classification: Further analysis
References
References
[AA1] Anton Antonov, "Handwritten digits recognition by matrix factorization", (2016), MathematicaForPrediction at WordPress.
[AA2] Anton Antonov, "Comparison of PCA and NNMF over image de-noising", (2016), MathematicaForPrediction at WordPress.
[AA3] Anton Antonov, "Comparison of PCA, NNMF, and ICA over image de-noising", (2016), MathematicaForPrediction at WordPress.
Cite this as: Anton Antonov, "Simplified Machine-Learning Workflow #10" from the Notebook Archive (2020), https://notebookarchive.org/2020-09-55swtlo
Download
