nipunbatra

(base) ➜  prog-ml.github.io git:(main) ✗ pip install black[jupyter]
zsh: no matches found: black[jupyter]
(base) ➜  prog-ml.github.io git:(main) ✗ black notebooks/introduction/simpson.ipynb 
Skipping .ipynb files as Jupyter dependencies are not installed.
You can fix this by running ``pip install black[jupyter]``
No Python files are present to be formatted. Nothing to do 😴
(base) ➜  prog-ml.github.io git:(main) ✗ which black
/Users/nipun/miniconda3/bin/black
(base) ➜  prog-ml.github.io git:(main) ✗ pip install 'black[jupyter]'              

nipunbatra

Before PR	After PR

nipunbatra

nipunbatra

nipunbatra

from nilmtk import DataSet
ds = DataSet("/home/nipunbatra-pc/Downloads/iawe.h5")

elec = ds.buildings[1].elec

elec.meters[0].load(chunksize=10)
next(elec.meters[0].load(chunksize=10))

nipunbatra

nipunbatra

#!/usr/bin/env python
# -*- coding: utf-8 -*-
#
#  pegasos.py
#  
#  Copyright 2013 nipun batra <nipunb@iiitd.ac.in>
#  
#  This program is free software; you can redistribute it and/or modify
#  it under the terms of the GNU General Public License as published by
#  the Free Software Foundation; either version 2 of the License, or

nipunbatra

ps aux|grep sparse-transfer-cv-static-matrix.py|awk '{print $2}'|xargs kill -9

nipunbatra

"""
========================================
Unfair Casino Problem using Discrete HMM
========================================

This script shows how to use Decoding in a Discrete(Multinomial) HMM, which means
given the model parameters and an observed sequence we wish to find the most likely hidden state sequence
generating the same.
It uses the model given in http://www.rose-hulman.edu/~shibberu/MA490/MA490HMM.html#Example_Dishonest_Casino_
One may also refer to a great lecture series at http://vimeo.com/7175217

nipunbatra

We've all seen those weird looking mathematics equations that pop up when we hear about Fourier transforms. In this blog post, we'll programatically try and develop an intuitive understanding into the whole process. While I'll be using the scientific Python stack in this blog post, code in Matlab, R should not be that different.

Goal: Given a time series signal, we'll try and decompose it into various contituent freqencies.

First, let us assume that we are doing some signal acquisition and we can sample at 100 Hz frequency (100 times per second). We collect data for 10 seconds. So, we have a total of 1000 samples.

Samples collection duration (T) = 10s
Sampling frequency (Fs) =100Hz