Global AI and Data Science

Global AI & Data Science

Train, tune and distribute models with generative AI and machine learning capabilities

 View Only

  • 1.  How to get good accuracy of a linear regression model having very few observations?

    Posted Mon June 22, 2020 04:56 PM
    I have to create a linear regression model for the collection forecasting. The data is grouping of monthly collections, about 35 observations. I've tried with lasso and ridge regression but the RMSE value is coming around 2%. what are the ways for optimizing my model so that RMSE value gets reduced.

    ------------------------------
    Kapil Kumar
    ------------------------------



    #AIandDSSkills
    #AIandDSSkills
    #DataandAILearning


  • 2.  RE: How to get good accuracy of a linear regression model having very few observations?

    Posted Tue June 23, 2020 03:22 AM
    Kapil,

    I think part of the question is to know what is the goal of your study?
    Given there's a time factor, I assume you want to predict future values, in which case it would probably be more suitable to use some TimeSeries prediction?

    ------------------------------
    Philippe Gregoire
    IBM France - TSP & ISV Technical Enablement - Data&AI, IoT Europe
    NICE, France
    ------------------------------

    Original Message


  • 3.  RE: How to get good accuracy of a linear regression model having very few observations?

    Posted Tue June 23, 2020 08:13 AM
    Hi Kapil,
    I am not sure I understand.  You have monthly collections.  Do you have 35 monthly collections of multiple data points OR is it 35 months of a single data point per month?  If the latter, do you have sufficient data points for a reliable regression model?  If the former, is it possible to lump all data points together or is there time series effect?

    ------------------------------
    Murray Woods
    Column 72
    ------------------------------

    Original Message