@incollection{fdi:010085545,
  title = {{A} framework for statistically-sound customer segment search},
  author = {{A}mer-{Y}ahia, {S}.  and {B}erti-{E}quille, {L}aure and {C}hibah, {A}.},
  editor = {},
  language = {{ENG}},
  abstract = {{W}e develop {S}4, a {S}tatistically-{S}ound {S}egment {S}earch framework that combines principled data partitioning and sound statistical testing to verify common hypotheses in retail data and return interpretable customer data segments. {O}ur framework accommodates one-sample, two-sample, and multiple-sample testing, to provide various aggregations and comparisons of customer transactions. {T}o control the proportion of false discoveries in multiple hypothesis testing, we enforce an {FDR}-controlling procedure and formulate a unified optimization problem that returns customer data segments that satisfy the test for a given significance level, maximize coverage of the input data, and are within a risk capital. {W}e develop a greedy algorithm to explore different data partitions and test multiple hypotheses in a sound manner. {O}ur extensive experiments on four retail data sets examine the interaction between significance, risk and coverage, and demonstrate the expressivity, usefulness, and scalability of {S}4 in practice.},
  keywords = {},
  booktitle = {2021 {IEEE} 8th {I}nternational {C}onference on {D}ata {S}cience and {A}dvanced {A}nalytics ({DSAA})},
  numero = {},
  pages = {1--10},
  address = {{P}iscataway},
  publisher = {{IEEE}},
  series = {},
  year = {2021},
  DOI = {10.1109/{DSAA}53316.2021.9564199},
  ISBN = {978-1-6654-2100-3},
  URL = {https://www.documentation.ird.fr/hor/fdi:010085545},
}