After going through the process of mining and selecting relevant patents (you can read all about it here), and ending up with 530 “Must Read” patents and 254 “Maybe” patents, I started the process of reading and summarizing them.
So far, in this phase, I have explored three patents:
- US9740752B2 – Determining User Personality Characteristics from Social Networking System Communications and Characteristics
- US9251468B2 – Inferring user profile attributes from social information
- US8583471B1 – Inferring household income for users of a Social Networking System
For ease of consultation and grasp of the information acquired, I have synthesized these patents further in three schematics (you can click on them to consult them in a larger size).
Each schematic has a patent classification and a data mining method. The patent classification was extracted from the official patent documentation, whilst the data mining method was borrowed from Koh Hian & Chan Kin Leong’s Data Mining and Customer Relationship Marketing in the Banking Industry, quoted in Christl’s and Silverman’s Networks of Control, according to which “data mining methods are classified according to the purpose they serve”:
- Methods for description and visualization
- Methods for association and clustering
- Methods for classification and estimation (prediction)
US9740752B2 – Determining User Personality Characteristics from Social Networking System Communications and Characteristics
US9251468B2 – Inferring user profile attributes from social information
US8583471B1 – Inferring household income for users of a Social Networking System
[…] but also annotating images, network analysis, genetic data). In the context of this research, given that LDA is possibly a method used for Facebook in clustering, our interest lies in understanding how it works, and which elements are […]