SUSHI Preference Data Sets

Bookmark this on Hatena Bookmark
Pocket

The SUSHI Preference Data Set includes responses of a questionnaire survey of preference in SUSHI. These preference are collected by a scoring method using a five-point-scale, and additionally by a ranking method. A ranking method is a one of method for performing a sensory test. In this method, the respondents sort given objects according to their preference order. This data set also includes demographic data of respondents and features of SUSHI.

Data Sets

The readme files ended in “-jp.txt” are written in Japanese, and those ended in “-en.txt” are written in English.

All Data Set

This data set includes all preference data, demographic data of respondents, and features of SUSHI. This is used for performing clustering in our publication [Mining Complex Data].

WARNING: The file downloaded before Dec. 29, 2015, includes a bug in data collected by a scoring method (i.e., a “*.score” file). In a readme file, I noted that the score “4” means “the most preferred.”, but it truly means “the disliked”. Currently, this bug is fixed. Note that no problem has been found for data collected by a ranking method (i.e., “*.order” files).

SUSHI Preference data set in the PrefLib format

This archive contains the sushi data set in the PrefLib format. You can find the details about this archive in the PrefLib site.

Collaborative Filtering Data Set

This data set is generated from the “All Data Set,” and is tailored for collaborative filtering tests. This data set is used in our publication [KDD03, RecSys10].

Object Ranking Data Set

This data set is generated from the “All Data Set,” and is tailored for object ranking tests. The file “sushi3.idata” in the above “sushi3.tgz” is also required to get the features of objects. This data set is used in our publication [Preference Learning].

License

  • We involve Toshihiro Kamishima, his colleagues, and their employers.
  • You involve the user of this data and his/her colleagues, and their employers.
  • We are NOT liable for any damages or losses, arising out of or related to your use or inability to use this data set.
  • You can use this data set for any research purpose.
  • You must not redistribute.
  • We would appreciated if you acknowledge the use of these program codes or data sets in publications by citing one of our related publications.

History

  • sushi1: contains 52 orders
  • sushi2: contains 1025 orders
  • sushi3: contains 5000 orders
  • sushi3-2016: contains 5000 orders, bug fix for score data