Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
This work was supported by the FP7-ICT project CARRE (No. 611140), funded in part by the
European Commission.
George Drosa...
G. Drosatos, IUPESM – World Congress 2015: 2
e-Health Application
e.g.:
 Personal
 Clinical
intentions,
plans,
beliefs,
...
G. Drosatos, IUPESM – World Congress 2015: 3
 It is a EU funded project in the area of cardiorenal with focus to
provide ...
G. Drosatos, IUPESM – World Congress 2015: 4
‒ Capture Personal Information
◦ Goal: Detect intentions
‒ Possible Sources:
...
G. Drosatos, IUPESM – World Congress 2015: 5
 What is privacy?
◦ “The right to be let alone” [Warren and Brandeis, 1890]
...
G. Drosatos, IUPESM – World Congress 2015: 6
main principle: preserve the patients’ privacy
G. Drosatos, IUPESM – World Congress 2015: 7
(1) Offline, initialization process
◦ predefine query categories (~250)
◦ cre...
G. Drosatos, IUPESM – World Congress 2015: 8
‒ Query Detector as a browser extension
◦ Firefox
◦ Chrome
‒ User Intention E...
G. Drosatos, IUPESM – World Congress 2015: 9
‒ Query Detector as a browser extension
◦ Firefox
◦ Chrome
‒ User Intention E...
G. Drosatos, IUPESM – World Congress 2015: 10
‒ Query Detector as a browser extension
◦ Firefox
◦ Chrome
‒ User Intention ...
G. Drosatos, IUPESM – World Congress 2015: 11
G. Drosatos, IUPESM – World Congress 2015: 12
 Conclusions
◦ Provide a proof of concept
◦ Apply a privacy by design appro...
G. Drosatos, IUPESM – World Congress 2015: 13
 Slides & Reprints: http://www.drosatos.info
 Online Demonstration: http:/...
G. Drosatos, IUPESM – World Congress 2015: 14
This work was supported by the FP7-ICT project CARRE (No. 611140),
funded in...
Upcoming SlideShare
Loading in …5
×

Extracting Intention from Web Queries– Application in eHealth Personalization

368 views

Published on

G. Drosatos, A. Arampatzis and E. Kaldoudi, Extracting Intention from Web Queries - Application in eHealth Personalization, In the Proceedings of the IUPESM 2015: World Congress on Medical Physics & Biomedical Engineering, Toronto, Canada, 7-12 June, 2015.

Personalizing healthcare applications requires capturing patient specific information, including medical history, health status, and mental aspects such as behaviors, intentions, and attitudes. This paper presents a privacy-friendly system to deduce patient intentions that can be used to personalized eHealth applications. In the proposed approach patient inten-tion is deduced from web query logs via query categorization techniques. The architecture assumes a user application which conceals the user’s queries from the central system, while only relevant intentions are disclosed. The paper presents a prototype implementation of the proposed architecture to extract intentions for personalizing empowerment services for the cardiorenal patient. Emphasis is placed on identifying intentions related to travel, diet and physical exercise, as these play an important role for the daily management of cardiorenal disease.

Published in: Science
  • Be the first to comment

  • Be the first to like this

Extracting Intention from Web Queries– Application in eHealth Personalization

  1. 1. This work was supported by the FP7-ICT project CARRE (No. 611140), funded in part by the European Commission. George Drosatos Avi Arampatzis and Eleni Kaldoudi School of Medicine Dept. of Electric and Computer Engineering Democritus University of Thrace
  2. 2. G. Drosatos, IUPESM – World Congress 2015: 2 e-Health Application e.g.:  Personal  Clinical intentions, plans, beliefs, etc. mental personal health records personal sensors physical electronic health records health insurance financial Personal Data Privacy
  3. 3. G. Drosatos, IUPESM – World Congress 2015: 3  It is a EU funded project in the area of cardiorenal with focus to provide personalized health  Personal data: Sensor data (e.g. activity and blood pressure), PHR and patient’s intentions (travel, diet, diseases, etc)
  4. 4. G. Drosatos, IUPESM – World Congress 2015: 4 ‒ Capture Personal Information ◦ Goal: Detect intentions ‒ Possible Sources: ◦ Social Media: Facebook, Twitter, etc ◦ Browsing History ◦ Web Searches ‒ First choice: Web searches to extract intentions ◦ Good source to reveal user’s interests and intentions ◦ Web search engines are one of the most popular uses of the web, e.g. >70% of internet users report looking online for health information *The Pew Research Center. (2013) Health Online 2013. http://www.pewinternet.org/2013/01/15/health-online-2013/
  5. 5. G. Drosatos, IUPESM – World Congress 2015: 5  What is privacy? ◦ “The right to be let alone” [Warren and Brandeis, 1890] ◦ “The right of the individual to decide what information about himself should be communicated to others and under what circumstances” [Westin, 1970] ◦ The right to informational self-determination [1983]  Personal Data: Any information that refers to a person  Related Legislation: e.g. EU Data Protection Directive 95/46/EC ◦ Indicative principles:  Reported and transparent processing  Finality & Purpose Limitation  Personal data quality  Security  Personal data traffic outside EU
  6. 6. G. Drosatos, IUPESM – World Congress 2015: 6 main principle: preserve the patients’ privacy
  7. 7. G. Drosatos, IUPESM – World Congress 2015: 7 (1) Offline, initialization process ◦ predefine query categories (~250) ◦ create index of documents = the collection of top most related documents to each category from a set representative of the entire web (ClueWeb 09_b) (2) Real-time repetitive process ◦ run user query in the index ◦ based on the results, associate user query with predefined categories *Agrawal R., Yu X et al. , Neural Information Proc., Springer, 7064 of LNCS, pp 148-157, 2011 privacy preserving: step #2 process is performed on user-side
  8. 8. G. Drosatos, IUPESM – World Congress 2015: 8 ‒ Query Detector as a browser extension ◦ Firefox ◦ Chrome ‒ User Intention Extractor as a Java application ◦ Platform independent
  9. 9. G. Drosatos, IUPESM – World Congress 2015: 9 ‒ Query Detector as a browser extension ◦ Firefox ◦ Chrome ‒ User Intention Extractor as a Java application ◦ Platform independent
  10. 10. G. Drosatos, IUPESM – World Congress 2015: 10 ‒ Query Detector as a browser extension ◦ Firefox ◦ Chrome ‒ User Intention Extractor as a Java application ◦ Platform independent
  11. 11. G. Drosatos, IUPESM – World Congress 2015: 11
  12. 12. G. Drosatos, IUPESM – World Congress 2015: 12  Conclusions ◦ Provide a proof of concept ◦ Apply a privacy by design approach in our methodology  Work in Progress ◦ Improve the technique of query classification ◦ Determine the safe detected intentions based on classification technique (without a fixed limit, e.g. n=3) ◦ Perform a user study in the side of patients in order to determine the correctness of intentions  Future Work ◦ Detect intentions from other online activities (e.g. social media) of patient ◦ Investigate how to utilize the intentions in a Decision Support System (DSS)
  13. 13. G. Drosatos, IUPESM – World Congress 2015: 13  Slides & Reprints: http://www.drosatos.info  Online Demonstration: http://youtu.be/IMHlIbwcDRY  You can find binaries and source codes at: https://www.carre-project.eu/innovation/web-lifestyle-data-aggregator/
  14. 14. G. Drosatos, IUPESM – World Congress 2015: 14 This work was supported by the FP7-ICT project CARRE (No. 611140), funded in part by the European Commission. CARRE Project: Personalized patient empowerment and shared decision support for cardiorenal disease and comorbidities.

×