2018 5th Learning Seminar “Derivation and Parameter Estimation of MNL”
The introduction has been completed up to the 4th seminar, and in the 5th and subsequent basic seminars, students will learn some specialized contents along with programming. This time it was “Derivation and Parameter Estimation of MNL”. Mr. Koizumi (M1) and Mr. Muro (M1) from Asakura Laboratory (a laboratory belonging to TSU) were in charge of this seminar.
The main feature was the introduction of the “maxLik” package in R statistical software. The “Optim” package used in the past had a problem of producing erroneous estimation results in transportation mode choice models when the sample to be analyzed differed depending on the mode of transportation. The new “maxLik” package was introduced to deal with this problem.