We are currently working on the following:
- a SIMALTO simulator; we are in an early exploratory phase – we are trying to define what features this simulator should have. If you would like to contribute with your idea, please get in touch with us firstname.lastname@example.org
Most important recent development work:
- in Choice-Based Conjoint (CBC), available in R-sw Conjoint, one or more fixed products can now be included in the exercise and utilities estimation process. Fixed products are those whose definitions do not vary across the CBC tasks (e.g., typically competitive products). The amended functions are available from R-sw 2022r3.
- Shapley Values (SV) algorithms available in R-sw Drivers to run derived importance analysis are now more powerful, allowing the analysis of data where the dependent variable is either on a rating / ratio scale or dichotomous (0/1 values). We implemented logistic SV in addition to our historical SV analysis based on a linear model. The additional functions (derived.importance.v2; derived.importance.v2.manyvar) are available from R-sw 2022r2.
- the TURF algorithm in R-sw Discriminant is now more powerful, allowing to take into account frequency of purchase (if available). The user can now force one or more items into the optimal combinations (if required). It is also possible to take into account the cost of individual items to identify optimal combinations within a predefined total cost range. A number of examples have been added to the package. The revised function (turf.analysis) and amended Excel simulator are available from R-sw 2020r2.
- R-sw Discriminant now includes a Genetic Algorithm to identify the combination(s) of variables able to most accurately predict cluster membership. This Genetic algorithm is very powerful, fast and more efficient versus traditional approaches, such as the one implemented in the function DA.report of R-sw Discriminant. The new function (genetic.DA) is available from R-sw 2019r2.
What are Genetic Algorithms? Genetic Algorithms are commonly used to generate good-quality solutions to difficult optimization and search problems by relying on genetics-inspired operators such as selection, crossover and mutation. A Genetic Algorithm repeatedly modifies a population of individual solutions; at each step, the algorithm selects individuals from the current population to be parents and uses them to produce the offsprings, subject to potential mutations, for the next generation. Over successive generations, the population evolves toward better solutions.
- In 2018 we have decided to re-write most of the functions available for R-sw Conjoint, in particular those for Rank-Based Conjoint and Allocation-Based Conjoint. The new functions are much simpler to use than previous ones, and allow running any utility estimation in a few minutes (literally 2 minutes to prepare the script). The new functions are available from R-sw 2019r1.
[last update: September 2022]