176985602337900

Find freelancers. Lose those costly delays.

With 1.7 million freelancers, we'll match you with the perfect talent.

or, Register with Elance »

Machine Learning or signal Processing Problem using Electric Meter Data
Sign in to Add to Watch List

IT & Programming > Data Science

View Change History

The client has made the following changes to the job.

Description
Date

Job description changed. New Version|Previous Version

Mar 6, 2014

Job title changed to "Machine Learning or signal Processing Problem using Electric Meter Data". Previously: "Machine Learning Problem using Electric Meter Data"

Skills updated.

Mar 6, 2014

Job description changed. New Version|Previous Version

Mar 6, 2014

Job description changed. New Version|Previous Version

Skills updated.

Mar 6, 2014
Close

Job Description

  |  Change History >>

Updated

Hi, I am trying to build a system where using the home electric meter (measurements every hour) and a characterization of high energy consumption equipment (HVAC, Heater, Dryer), I can guess when appliances were turned on and off and how long they were used for. Naturally there will be a margin of error but I believe with a rich training set, this can be reduced. Note that the challenge is trying to match the presence of multiple signals (appliance characterization) within the measurement data. For example, the appliance characterization of a electric heater could be a signal that has a high power consumption for some duration t and then turns off (when thermostat deactivates) and then turns on again. I would like to get a proposal of how you would address this problem and what type of model you would use (e.g. naive bayesian, neural network, etc) and what tools you would use to solve this problem (e.g. matlab, R). The output should be the data stating which appliance for which time windows. The first step would be to simply build the concept engine solution of taking the training set and defining the model. And then trying it out on real data that I have to see how good its performance is. Next step would be to build the machine learning engine where I can feed it future training sets, additional appliance characterization models and visualization for the engine's performance. The system should be as automated as possible so that its simple to improve the engine and use the engine in batch with the raw data that would come in. Attached is the raw measurement and timestamp data. This is not the training set. I have training set but I want to see what your proposed approach would be so that I know that you have the capability to handle such a project. Thanks. If things go well, this could be one of many projects that we work on together. sri

Hi, I am trying to build a system where using the home electric meter (measurements every hour) and a characterization of high energy consumption equipment (HVAC, Heater, Dryer), I can guess when appliances were turned on and off and how long they were used for. Naturally there will be a margin of error but I believe with a rich training set, this can be reduced. THe main issue is to find when an appliance was turned on and how long the appliance was active for by only looking at a set of discrete electric meter measurements. Note that the challenge is trying to match the presence of multiple signals (appliance characterization) within the measurement data. For example, the appliance characterization of a electric heater could be a signal that has a high power consumption for some duration t and then turns off (when thermostat deactivates) and then turns on again. I would like to get a proposal of how you would address this problem and what type of model you would use (e.g. naiv...

Mar 6, 2014

Hi, I am trying to build a system where using the home electric meter (measurements every hour) and a characterization of high energy consumption equipment (HVAC, Heater, Dryer), I can guess when appliances were turned on and off and how long they were used for. Naturally there will be a margin of error but I believe with a rich training set, this can be reduced. I would like to get a proposal of how you would address this problem and what type of model you would use (e.g. naive bayesian, neural network, etc) and what tools you would use to solve this problem (e.g. matlab, R). The output should be the data stating which appliance for which time windows. The first step would be to simply build the concept engine solution of taking the training set and defining the model. And then trying it out on real data that I have to see how good its performance is. Next step would be to build the machine learning engine where I can feed it future training sets, additional appliance characterization models and visualization for the engine's performance. The system should be as automated as possible so that its simple to improve the engine and use the engine in batch with the raw data that would come in. Attached is the raw measurement and timestamp data. This is not the training set. I have training set but I want to see what your proposed approach would be so that I know that you have the capability to handle such a project. Thanks. If things go well, this could be one of many projects that we work on together. sri

Hi, I am trying to build a system where using the home electric meter (measurements every hour) and a characterization of high energy consumption equipment (HVAC, Heater, Dryer), I can guess when appliances were turned on and off and how long they were used for. Naturally there will be a margin of error but I believe with a rich training set, this can be reduced. Note that the challenge is trying to match the presence of multiple signals (appliance characterization) within the measurement data. For example, the appliance characterization of a electric heater could be a signal that has a high power consumption for some duration t and then turns off (when thermostat deactivates) and then turns on again. I would like to get a proposal of how you would address this problem and what type of model you would use (e.g. naive bayesian, neural network, etc) and what tools you would use to solve this problem (e.g. matlab, R). The output should be the data stating which appliance for which time windows. The first step would be to simply build the concept engine solution of taking the training set and defining the model. And then trying it out on real data that I have to see how good its performance is. Next step would be to build the machine learning engine where I can feed it future training sets, additional appliance characterization models and visualization for the engine's performance. The system should be as automated as possible so that its simple to improve the engine and use the engine in batch with the raw data that would come in. Attached is the raw measurement and timestamp data. This is not the training set. I have training set but I want to see what your proposed approach would be so that I know that you have the capability to handle such a project. Thanks. If things go well, this could be one of many projects that we work on together. sri

Mar 6, 2014

Hi, I am trying to build a system where using the home electric meter (measurements every hour) and a characterization of high energy consumption equipment (HVAC, Heater, Dryer), I can guess when appliances were turned on and off and how long they were used for. Naturally there will be a margin of error but I believe with a rich training set, this can be reduced. I would like to get a proposal of how you would address this problem and what type of model you would use (e.g. naive bayesian, neural network, etc) and what tools you would use to solve this problem (e.g. matlab, R). The output should be the data stating which appliance for which time windows. Attached is the raw measurement and timestamp data. This is not the training set. I have training set but I want to see what your proposed approach would be so that I know that you have the capability to handle such a project. Thanks. If things go well, this could be one of many projects that we work on together. sri

Hi, I am trying to build a system where using the home electric meter (measurements every hour) and a characterization of high energy consumption equipment (HVAC, Heater, Dryer), I can guess when appliances were turned on and off and how long they were used for. Naturally there will be a margin of error but I believe with a rich training set, this can be reduced. I would like to get a proposal of how you would address this problem and what type of model you would use (e.g. naive bayesian, neural network, etc) and what tools you would use to solve this problem (e.g. matlab, R). The output should be the data stating which appliance for which time windows. The first step would be to simply build the concept engine solution of taking the training set and defining the model. And then trying it out on real data that I have to see how good its performance is. Next step would be to build the machine learning engine where I can feed it future training sets, additional appliance characterization models and visualization for the engine's performance. The system should be as automated as possible so that its simple to improve the engine and use the engine in batch with the raw data that would come in. Attached is the raw measurement and timestamp data. This is not the training set. I have training set but I want to see what your proposed approach would be so that I know that you have the capability to handle such a project. Thanks. If things go well, this could be one of many projects that we work on together. sri

Mar 6, 2014

Close
  • Posted: Thu, Mar 06, 2014
  • Time Left: Closed
  • Location: Anywhere
  • Client prefers freelancers from:
    Anywhere

    You are still able to submit a proposal for this job.

    The client prefers freelancers from
    a different location.

    You're still able to submit a proposal for this job, regardless of your location.
  • Start: Immediately
  • Hourly Rate: Not Sure
  • Hrs/wk: Not Sure | Duration: Not Sure
  • Work View™ Payment Protection
  • U.S. freelancers must have W9
Sign in to view client's details

Hi,
I am trying to build a system where using the home electric meter (measurements every hour) and a characterization of high energy consumption equipment (HVAC, Heater, Dryer), I can guess when appliances were turned on and off and how long they were used for. Naturally there will be a margin of error but I believe with a rich training set, this can be reduced.

THe main issue is to find when an appliance was turned on and how long the appliance was active for by only looking at a set of discrete electric meter measurements.

Note that the challenge is trying to match the presence of multiple signals (appliance characterization) within the measurement data. For example, the appliance characterization of a electric heater could be a signal that has a high power consumption for some duration t and then turns off (when thermostat deactivates) and then turns on again.

I would like to get a proposal of how you would address this problem and what type of model you would use (e.g. naiv...

Sign in or Register to see more

Job ID: 53978036
Proposals
Hourly Rate: Avg $ | High $ | Low $ — Show Pricing
  • Submit Date (Latest)

Ivan M.      
Ivan M.'s identity has been verified.  Learn More
 Russia  |  
PhD, Moscow State university of railway engineering, specialization in Automatic Control and Informatics in Technical Systems. At the moment I'm...
2
  |  
 0.0   |  $0 Earnings   |  0 Jobs
Bid ID: 54506867  |  Submitted: Mar 19, 2014 14:03 ET 
Proposal SEALED

 United States  |  
Seasoned engineer with over 20 years of experience in complex algorithm development for signal and image processing applications. Have worked in...
6
  |  
 5.0   |  $13,158 Earnings   |  2 Jobs
Bid ID: 54434794  |  Submitted: Mar 17, 2014 22:19 ET 
Proposal SEALED

 Philippines  |  
Will provide creative and innovative solutions to your needs!
1
  |  
 0.0   |  Private   |  0 Jobs
Bid ID: 54251231  |  Submitted: Mar 12, 2014 19:44 ET 
Proposal SEALED

 Belgium  |  
Doing projects in DSP, Image Processing, Computer Vision, Pattern Recognition, mathematical algorithm implementation etc. Working since 2005 on...
1
  |  
 0.0   |  $0 Earnings   |  0 Jobs
Bid ID: 54228335  |  Submitted: Mar 12, 2014 10:32 ET 
Proposal SEALED

 India  |  
HIRE EXPERTS FROM TECHSTORM TO FIRE YOUR IMAGINATION! TechStorm IT Solutions provides innovative value added solutions for Image & Video...
4
  |  
 0.0   |  $0 Earnings   |  0 Jobs
Bid ID: 54167632  |  Submitted: Mar 11, 2014 08:06 ET 
Proposal SEALED

 Italy
I'm an engineer with experience in developing cross platform computer vision applications, 2D and 3D image processing and analisys with machine...
3
  |  
 4.6   |  Private   |  4 Jobs
Bid ID: 54110667  |  Submitted: Mar 10, 2014 07:24 ET 
Proposal SEALED

 Germany  |  
INVITED
R&D experience in various fields
1
  |  
 4.9   |  $5,443 Earnings   |  8 Jobs
Bid ID: 54087987  |  Submitted: Mar 9, 2014 18:06 ET 
Proposal SEALED

 United States  |  
http://www.autopredictivecoding.com/ is a Saas based prediction engine that generates automatic prediction using advanced machine learning...
1
  |  
 0.0   |  Private   |  0 Jobs
Bid ID: 54079747  |  Submitted: Mar 9, 2014 12:30 ET 
Proposal SEALED

 United States  |  
I hold a PhD in electrical and computer engineering. I have over 10 years working experience in the field of machine learning, data mining, pattern...
Skills: Data Modeling, C, Java
1
  |  
 0.0   |  $0 Earnings   |  0 Jobs
Bid ID: 54066247  |  Submitted: Mar 8, 2014 23:17 ET 
Proposal SEALED

Big Data Power (Hadoop, Pytho...      
One or more team members at  are verified.  Learn More
 India  |  
Our expertise includes Big Data, Databases, Search, Machine Learning and NLP applications. We have experience in following areas: * Hadoop -...
12
  |  
 5.0   |  Private   |  3 Jobs
Bid ID: 54048559  |  Submitted: Mar 8, 2014 07:54 ET 
Proposal SEALED

 United States  |  
Embedded Systems Developers
1
  |  
 0.0   |  $0 Earnings   |  0 Jobs
Bid ID: 54035899  |  Submitted: Mar 7, 2014 21:37 ET 
Proposal SEALED

 Spain  |  
INVITED
I am an engineer and software developer with around 14 years experience. My strong suits are C/C++/C#/Ruby and scientific computing.
3
  |  
 5.0   |  $5,143 Earnings   |  4 Jobs
Bid ID: 54022643  |  Submitted: Mar 7, 2014 13:09 ET 
Proposal SEALED
Sign in to Elance and start working on jobs today.
Sign in to view more of the job details and submit a proposal. Once registered, you'll have access to thousands of jobs online or through email.
Are you ready to post a job like this one?
Post a Similar Job »