search

cran / MCI

:exclamation: This is a read-only mirror of the CRAN R package repository. MCI — Multiplicative Competitive Interaction (MCI) Model
0 stars 0 forks source link

Market Shares vs. Patronage Probability #1

Open nelsonSchwarz opened 7 years ago

nelsonSchwarz commented 7 years ago

Hi,

It is my understanding that the results of the Huff model are probabilities for the decision of customers to shop at a location, given a competitive environment (interaction probabilities). And these probabilities can be interpreted as market shares of the regarded locations (j) in the customer origins (i): p_ij (local market shares).

However, the results of the huff.shares() function (i.e. p_ij) are solely market share percentages. How can I get the patronage probability, not the market share?

Evidently the market share is not the probability, as p_ij is not 0 < p_ij < 1 and the cumulative sum of p_ij is always the total number of origins. Would this be the right way:

huffworkfile$U_ij / sum(huffworkfile$U_ij)

If so, what is happening in line 60? Should this part not be calculating the actual probability?

huffworkfile$p_ij <- as.numeric(huffworkfile$U_ij)/as.numeric(huffworkfile$sum_U_ij)

Please clarify. Thank you.

Best, Nelson

gaborcsardi commented 7 years ago

Hi, this is a read-only mirror of CRAN, please send your requests to the original package authors, see the DESCRIPTION file.