CMN 170 Lecture Notes - Lecture 7: Mobile Phone, Big Data, Unsupervised Learning
Document Summary
Empirical: big data based on what happened in the past. Theoretical: computer simulations about what could happen in the future. Biggest stumbling block is obtaining the data to parameterize and validate . Digital footprint, lots of data (n=n), data fusion (information is unstructured and incomplete), often available in real-time (it"s dynamic), made machine learning algorithms effective (no need for theory) Allows people to learn about individual and behavior patterns. Can use data records like call duration and call frequency to predict socio-economic, demographic and other behavioral traits. Data fusion complements and fills up holes of missing data. Big data paradigm: looking for many different sources and bringing them together to get one coherent outlook = data fusion. Mobile phone operators sell mobility data to business owners to optimize for gain. Consumers" financial vulnerability: social influencer: product spreading, rural and barely making it, ethnic second-city strugglers, retiring on empty : singles, tough start: young single parents, credit crunched: city families.
