+ Reply to Thread
Results 1 to 2 of 2

Thread: Math Help: How to Calculate Weights for Multiplicative Model with Dummy Variables

  1. #1
    Points: 9, Level: 1
    Level completed: 17%, Points required for next Level: 41

    Posts
    2
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Math Help: How to Calculate Weights for Multiplicative Model with Dummy Variables




    Hi Everyone,

    My question is how do I calculate weights for Y=W1X1*W2X2*WNXN where N is the number of independent variables.

    I could use logarithms to convert this to an additive model, but each of the independent variables are binary/dummy variables. Here is some background.

    Assume two independent variables (more in practice). Really this is two sets of dummy variables. The predictive airline market share model looks something like this:

    ShareIndex=
    [W(ns)D(ns)+W(1s)D(1s)+W(2s)D(2s)] *
    [W(wb)D(wb)+W(nb)D(nb)+W(rj)D(rj)+W(tp)D(tp)]

    W for weights, D for binary dummy variables
    ITINERARY CONNECTIVITY (ns=nonstop, 1s=connection, 2s=doubleconnection)
    AIRCRAFT TYPE (wb=widebody, nb=narrowbody, rj=regionaljet, tp=turboprop)

    For each record, the there is one AIRCRAFTTYPE dummy variable with a value of 1; all others 0.

    For each record, the there is one ITINERARY dummy variable with a value of 1; all others 0.

    So basically the model is ShareIndex=W(aircrafttype)*W(itinerarytype)

    The dependent variable in the actual data is SHARE (market share). However, the model might produce a ShareIndex where each record's SHAREINDEX / (sum SHAREINDEXES all records that market)=market share

    For example if a market has 10 records each with 10% market share. MS(record1)=W(AircraftType)*W(ItineraryType)/

    How do I calculate the coefficients/weights?

    Thanks!!!

  2. #2
    Points: 9, Level: 1
    Level completed: 17%, Points required for next Level: 41

    Posts
    2
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: Math Help: How to Calculate Weights for Multiplicative Model with Dummy Variable


    Here is some more info...

    Looking for to calculate the weights/coefficients. Not sure how to regress this model. (This model is more complex than any I have worked with due to multiplicative function AND use of ShareIndex while the data only includes Market share.)

    MarketShare = 𝑆ℎ𝑎𝑟𝑒𝐼𝑛𝑑𝑒𝑥(&#119894 / (∑_(𝑖=1)^𝑛▒(𝑆ℎ𝑎𝑟𝑒𝐼𝑛𝑑𝑒𝑥(&#119894) )

    My data array looks something like this...

    i,Market,Share,Dns,D1s,D2s,Dwb,Dnb,Drj
    1,IAHORD, 0.205 ,1,0,0,0,1,0
    2,IAHORD, 0.158 ,1,0,0,0,1,0
    3,IAHORD, 0.299 ,1,0,0,0,1,0
    4,IAHORD, 0.181 ,1,0,0,0,1,0
    5,IAHORD, 0.157 ,1,0,0,0,1,0
    1,AUSGIG, 0.250 ,0,1,0,1,0,0
    2,AUSGIG, 0.219 ,0,1,0,1,0,0
    3,AUSGIG, 0.048 ,0,0,1,0,1,0
    4,AUSGIG, 0.020 ,0,0,1,0,1,0

    13,AUSGIG, 0.018 ,0,0,1,0,1,0

    Thanks again for your help.

+ Reply to Thread

           




Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats