Sunteți pe pagina 1din 3

Cold start

Cold start is a potential problem in computer-based information systems which involve a degree of automated data modelling.
Specifically, it concerns the issue that the system cannot draw any inferences for users or items about which it has not yet gathered
sufficient information.

Contents
1 Systems affected
2 Solutions
3 See also
4 References
5 External links

Systems affected
The cold start problem is most prevalent in recommender systems. Recommender systems form a specific type of information
filtering (IF) technique that attempts to present information items (movies, music, books, news, images, web pages) that are likely of
interest to the user. Typically, a recommender system compares the user's profile to some reference characteristics. These
characteristics may be from the information item (the content-based approach) or the user's social environment (the collaborative
filtering approach).

In the content-based approach, the system must be capable of matching the characteristics of an item against relevant features in the
user's profile. In order to do this, it must first construct a sufficiently-detailed model of the user's tastes and preferences through
preference elicitation. This may be done either explicitly (by querying the user) or implicitly (by observing the user's behaviour). In
both cases, the cold start problem would imply that the user has to dedicate an amount of effort using the system in its 'dumb' state
contributing to the construction of their user profile before the system can start providing any intelligent recommendations.

In the collaborative filtering approach, the recommender system would identify users who share the same preferences (e.g. rating
patterns) with the active user, and propose items which the like-minded users favoured (and the active user has not yet seen). Due to
[1]
the cold start problem, this approach would not consider items that no user had yet rated.

The cold start problem is also exhibited by interface agents. Since such an agent typically learns the user's preferences implicitly by
observing patterns in the user's behaviour "watching over the shoulder" it would take time before the agent may perform any
adaptations personalised to the user. Even then, its assistance would be limited to activities which it has formerly observed the user
engaging in.[2]

Solutions
There are several solutions that have been proposed to tackle the cold start problem. One of the effective solutions is to apply active
learning (machine learning)techniques, i.e., selectively choosing and obtaining more data, that can most improve the performance of
the recommender system. This is done by analysing the available data and estimating the usefulness of the data points (e.g.,
ratings).[3] In collaborative filtering recommender systems, these techniques are so called rating elicitation strategies.
[4][5]
In scenarios involving interface agents, the cold start problem may be overcome by introducing an element of collaboration amongst
agents assisting various users. This way, novel situations may be handled by requesting other agents to share what they have already
learnt from their respective users.[2]

In recommender systems, the cold start problem is often reduced by adopting a hybrid approach between content-based matching and
collaborative filtering. New items (which have not yet received any ratings from the community) would be assigned a rating
automatically, based on the ratings assigned by the community to other similar items. Item similarity would be determined according
to the items' content-based characteristics.[1]

The construction of the user's profile may be automated by integrating information from other user activities, such as browsing
histories. If, for example, a user has been reading information about a particular music artist from a media portal, then the associated
[6]
recommender system would automatically propose that artist's releases when the user visits the music store.

It is also possible to create initial profile of a user based on the personality characteristics of the user and use such profile to generate
personalized recommendation.[7][8] Personality characteristics of the user can be identified using a personality model such as five
factor model (FFM).

See also
Collaborative filtering
Preference elicitation
Recommender system
Active learning (machine learning)
Five Factor Model

References
1. Andrew I. Schein; Alexandrin Popescul; Lyle H. Ungar; David M. Pennock (2002).Methods and Metrics for Cold-
Start Recommendations(http://citeseer.ist.psu.edu/schein02methods.html). Proceedings of the 25th Annual
International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR 2002). New
York City, New York: ACM. pp. 253260. ISBN 1-58113-561-0. Retrieved 2008-02-02.
2. Yezdi Lashkari; Max Metral; Pattie Maes (1994). Collaborative Interface Agents(http://citeseer.ist.psu.edu/lashkari94
collaborative.html). Proceedings of the Twelfth National Conference on Artificial Intelligence.Seattle, Washington:
AAAI Press. pp. 444449. ISBN 0-262-61102-3. Retrieved 2008-02-02.
3. Rubens, Neil; Elahi, Mehdi; Sugiyama, Masashi; Kaplan, Dain (2016)."Active Learning in Recommender Systems"
(https://rd.springer.com/chapter/10.1007/978-1-4899-7637-6_24). In Ricci, Francesco; Rokach, Lior; Shapira,
Bracha. Recommender Systems Handbook(https://rd.springer.com/book/10.1007/978-1-4899-7637-6) (2nd ed.).
Springer US. ISBN 978-1-4899-7637-6.
4. Elahi, Mehdi; Ricci, Francesco; Rubens, Neil.Active Learning in Collaborative Filtering Recommender Systems(http
s://rd.springer.com/chapter/10.1007/978-3-319-10491-1_12). Springer International Publishing. pp. 113124.
ISBN 978-3-319-10491-1.
5. Elahi, Mehdi; Ricci, Francesco; Rubens, Neil (2016).A survey of active learning in collaborative filtering
recommender systems(https://www.researchgate.net/publication/303781992_A_survey_of_active_learning_in_colla
borative_filtering_recommender_systems). Computer Science Review via Elsevier .
6. "Vendor attempts to crack 'cold start' problemin content recommendations"(https://web.archive.org/web/200808010
00000/http://www.xiam.com/press/MM062907.pdf) (PDF). Mobile Media. United Kingdom: Informa Telecoms &
Media: 18. 2007-06-29. Archived fromthe original (http://www.xiam.com/press/MM062907.pdf)(PDF) on 2008-08-
01. Retrieved 2008-02-02.
7. Tkalcic, Marko; Chen, Li (2016)."Personality and Recommender Systems"(https://rd.springer.com/chapter/10.1007/
978-1-4899-7637-6_21). In Ricci, Francesco; Rokach, Lior; Shapira, Bracha.Recommender Systems Handbook(htt
ps://rd.springer.com/book/10.1007/978-1-4899-7637-6) (2nd ed.). Springer US.ISBN 978-1-4899-7637-6.
8. Fernndez-Tobas, Ignacio; Braunhofer, Matthias; Elahi, Mehdi; Ricci, Francesco; Cantad or, Ivn (2016). "Alleviating
the new user problem in collaborative filtering by exploiting personality information"
(https://rd.springer.com/article/1
0.1007%2Fs11257-016-9172-z). User Modeling and User-Adapted Interaction. doi:10.1007/s11257-016-9172-z(http
s://doi.org/10.1007%2Fs11257-016-9172-z).

External links
http://activeintelligence.org/wp-content/papercite-data/pdf/Rubens-Active-Learning-RecSysHB2010.pdf
http://activeintelligence.org/research/al-rs/

Retrieved from "https://en.wikipedia.org/w/index.php?title=Cold_start&oldid=802838023


"

This page was last edited on 28 September 2017, at 20:00.

Text is available under theCreative Commons Attribution-ShareAlike License ; additional terms may apply. By using this
site, you agree to the Terms of Use and Privacy Policy. Wikipedia is a registered trademark of theWikimedia
Foundation, Inc., a non-profit organization.

S-ar putea să vă placă și