Being Smart with Data
Autor: | H. Jost Reinhold, Anna Picco-Schwendener, Lorenzo Cantoni |
---|---|
Rok vydání: | 2019 |
Předmět: |
Information retrieval
Random assignment business.industry Computer science Decision tree learning 05 social sciences 020206 networking & telecommunications 02 engineering and technology Cross-validation Personalization Set (abstract data type) Moment (mathematics) Analytics 0502 economics and business 0202 electrical engineering electronic engineering information engineering 050211 marketing business Municipal wireless network |
Zdroj: | DG.O |
DOI: | 10.1145/3325112.3329884 |
Popis: | This study explores whether it is possible to automatically assign a user of a Municipal Wireless Network (MWN) to one of the previously identified user groups, using only the very small amount of data available at the moment of connection. To do so, researchers classified the users of the “WiFi Lugano” MWN into business travelers, leisure tourists, and locals in two different ways: 1) based on the users’ network usage, and 2) based on the users’ declared reason for being in Lugano. The two groupings allowed creating a two-tiered classification tree with cross validation resulting in a set of classification rules that would subsequently enable the system to assign future users to one of the three groups by using only a few variables known at the moment of connection. The identified rules make it possible to significantly increase the probability of a correct assignment over that of a random assignment. The study thus provides a general framework for automatically classifying users of a MWN with very limited available data, which can be applied also to other MWNs. Being able to distinguish users of a MWN at the moment of connection could allow the city to communicate in a more targeted way with each user group, for example through the use of tailored landing pages and thus to improve its relationship with citizens (G2C), businesses (G2B), and visitors (G2V). |
Databáze: | OpenAIRE |
Externí odkaz: |