GNU Octave Jobs

4 were found based on your criteria

  • Fixed-Price – Est. Budget: $300.00 Posted
    Hello, Machine Learning Specialist. I have a dataset of electricity generation data with weather dataset, and now I am developing machine learning model to predict month-ahead electricity generation. Criteria for Machine Learning Specialist is... * Can develop various models (multiple-variable regression, multiple-polynomial regression, Neural Network, SVM) * Compute cost-function * Use cross-validation and Regularization. * Octave is preferred. Data set is below. (not massive, just 1000 records) * Electricity generation (month) * Average Temperature (month) * Average Day Light ((month)) * Month (Jan- Dec) * Year (2010-2013) &&IMPORTANT&& If ...
  • Hourly – Less than 1 month – 10-30 hrs/week – Posted
    This work is to help with some research into an approach for smoothing time series data, using Loess smoothing and neural networks. This involves developing an approach that meets certain criteria that we will specify, experimenting and testing different techniques, documenting the algorithm developed and working with a systems analyst who will build it. We need someone with excellent mathematical skills and familiarity with an analysis tool such as Mathematica, Octave, NumPy and neural networks. The successful applicant will be ...
  • Hourly – Less than 1 month – 10-30 hrs/week – Posted
    The project is to develop a smoothing algorithm for real time data, extending and formalising some work that I have done. I need an applied mathematician with knowledge and experience in the following areas (with machine learning the most important): Data smoothing approaches, particularly Loess smoothing and Kalman filters Signal processing of times series data Machine learning algorithms, including k-nearest neighbour, Bayesian models, hidden Markov models Splines (MARS) Familiarity with tool sets, e.g. Python libraries, Octave, R or Mathematica ...
  • Fixed-Price – Est. Budget: $30.00 Posted
    I have a GNU Octave script that performs multivariate linear regression. I need the script modified. The error function algorithm for regression is normally: 1. actual - predicted, then square 2. sum all the individual errors 3. divide sum by number of observations 4. take square root Your Octave code that does this looks someting like this: SSerr = actual - predicted SSerr.^2 (SSerr/demo).^0.5 I have read that it is possible, and potentially useful, to perform regression using other ...
loading