hadoop-mapreduce-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Adaryl \"Bob\" Wakefield, MBA" <adaryl.wakefi...@hotmail.com>
Subject Re: PCA in HAdoop MapReduce
Date Wed, 30 Jul 2014 21:51:05 GMT
This may not be the right place to ask this question. I asked a more generic question about
how to do predictive modeling on hadoop and nobody answered. It perplexes me as well how to
take these machine learning concepts and implement them in a Map Reduce paradigm.

Adaryl "Bob" Wakefield, MBA
Mass Street Analytics

From: unmesha sreeveni 
Sent: Wednesday, July 30, 2014 1:27 AM
To: User Hadoop 
Subject: PCA in HAdoop MapReduce

​I am little bit confused in doing PCA 

I am trying it myown and I refered http://nyx-www.informatik.uni-bremen.de/664/1/smith_tr_02.pdf

1.MR job to calculate mean
2.MR Job to substract mean from the input data(Data Adjust in .pdf)
3.MR job to find covarience and Calculated the eigenvectors and eigenvalues of the covariance
Now need to transform my data
  Step 5: Deriving the new data set in pdf.
It is matrix multiplication ie MR3(step 3) * MR2(Step 2)

Matrix multiplication can be done in using a seperate mapreduce job but i have MR2 data but
I need to transpose MR3 data ie Final data = RowFeatureVector(MR3) * Row DataAdjust(MR2)

WHile doing transpose in another MR,we will not be sure about the order of elements they come
from mapper

How will we achieve the above transpose.

Please correct me if I am wrong in PCA concepts.​


Thanks & Regards 

Unmesha Sreeveni U.B

Hadoop, Bigdata Developer
Center for Cyber Security | Amrita Vishwa Vidyapeetham


View raw message