The code in this repository has been used to research the effect of linguistic and formality alignment on user satisfaction. This has been done by creating three versions of the already existing A.L.I.C.E. chatbot. Version 1 returns the best response based on linguistic alignment, version 2 is the current version of A.L.I.C.E. and version 3 returns the best response based on formality alignment.
After having downloaded the code, run the chatbots by typing the command ./alice#.sh in the following folder: ../ALICE#/bots/alice2
Replace # by the desired version number:
- Linguistic alignment
- Current
- Formality alignment
All tests and corresponding results that have been done to be able to find the best measures to accomplish this research can be found in the folder Measure_tests. The results of the 60 participants who have tested the 3 versions of the chatbot and rated them on user satisfaction can be found in the folder User_tests.
Disclaimer: The chatbot A.L.I.C.E. is created by Dr. Richard S. Wallace. All code for A.L.I.C.E. that has been used in this thesis has originally been written by Keith Sterling. However, the client#.py files (in the folder ../ALICE#/src/programy/clients/events/console) have been edited to a great extent to fit the purpose of my thesis.