L. Ardissono and A. Goy, `Tailoring the interaction with users in Web stores', User Modeling and User-Adapted Interaction, 10(4), 251--303, (2000).  Home/Search   Document Details and Download   Summary   Related Articles   Check

....cannot tell that she is not interested in blue telephones and red cars. Or, for that matter, the system cannot tell that the user is interested in large texts and small pictures, but not in small texts and large pictures. This problem gets even worse when more attributes and objects are involved. [1] suggest a similar approach: with probabilities it is dif cult to specify a property as simply irrelevant, and it is dif cult to deal with a lack of data: if we don t know yet about the users preferences wrt. some feature then what 2 Furthermore the technique of a Bayesian classi er is hard to ....

....a weighting of properties. Unfortunately, it is not clear how to learn these values, 3 and it shares the expressivity prob lem with naive Bayesian classi cation. 2There are a number smoothing methods for the probabilities that treat this problem, but their effectivity is domain dependent. 3In [1] the shop operator has to give the values for a number of user 5.4 Summary of the learning algorithms As it tums out, none of the learning algorithms we have investigated ful 11s all of the criteria established in Section 4. We currently use the algorithm CDL4 since it ful 11s most of the given ....

L. Ardissono and A. Goy. Tailoring the interaction with users in web stores. User Modeling and UserAdapted Interaction, 2001. (to appear).

....to assign the user to the best profile. Browsing starts from the presentation unit associated to a starting node. If the user is already registered, the last A(k) is set as current. Otherwise, he she is assigned to a generic profile, or to one calculated on the basis of a questionnaire (see [4] for an interesting way to interpret results in a probabilistic way) the initial value of A(k) is called A 0 (k) When the user visiting the node R r 1 requests to follow a link, the system computes the new PDF A(k) on the basis of the User Behaviour Variables and of s(k) see Section 2.3.1) ....

....by the terminal User Agent (browser) The Application Layer is the core of the system: it collects the user behaviour and characteristics and implements the adaptation process. It comprises two main modules: the Adaptive Hypermedia Application Server (AHAS) and the User Modelling Component (UMC) [4]; they run together with a Web Server. The UMC maintains the most recent actions of the user and executes the algorithm for the evaluation of the user s profile. After a user has selected the next page and the system has determined his her user s position in the Adaptation Space, the AHAS ....

Ardissono, L., Goy, A., Tailoring the Interaction With Users in Web Stores", inUser Modeling and UserAdapted Interaction, 10(4), Kluwer Academic Publishers, 2000.

....to the most recent hypotheses about her his knowledgeability and interests. Several aspects of the interaction can be customized: For instance, the layout of the interface, the amount of information to be displayed and the type of questions asked during the configuration of the product (see [1] for the definition of personalization strategies for a B2C application) In this description, we will focus on the last aspect, which strongly depends on the user s knowledgeability and interests and is critical to the usability of the configuration system. The basic idea is that the information ....

....models or knowledge exchange. However, recent trends show that extensions relying on the component port model for configuration from [8] as well as conceptual modeling techniques (e.g. based on the Unified Modeling Language) are useful in finding a unified view on the configuration problem ([1], 10] 12] This view on the configuration problem can also be found in commercial configuration tools [7] Our knowledge acquisition approach relies on the idea of using the Unified Modeling Language to model generic product models [1] because the notation is widely used in industrial ....

[Article contains additional citation context not shown here]

Ardissono L. and Goy A., Tailoring the Interaction With Users in Web stores. User Modeling and User-Adapted Interaction, 10(4), , Kluwer Academic Publishers, 2000, pp. 251-303.

....presented to the user depending on his or her expertise (Sales Assistant, Popp and Ldel, 1996) presenting expertise dependent explanations and technical details (Metadoc, Boyle and Encarnacion, 1994; KN AHS, Kobsa et al. 1994) and . generating expertise dependent product descriptions (SETA, Ardissono and Goy 1999, 2000b; Ardissono et al. 1999) Adaptation to a users knowledge of domain concepts, of rules and of other items is also a typical feature of intelligent tutoring systems. In many such systems, user knowledge is taken into account when guiding the user through the learning material. Examples are the ....

....based on concept relationships represented as domain knowledge. Figure 5 shows a graphical notation of a concept hierarchy (i.e. an ontology) from the animal kingdom, as it might be used for representing user knowledge about this domain (cf. Akoulchina and Ganascia, 1997; Milosavljevic, 1997; Ardissono and Goy, 1999, 2000b; Ardissono et al. 1999) 10 A fourth type, abductive reasoning (from the consequences to the premises) is also sometimes employed (e.g. in plan recognition, see Section 2.1.5) but will not be discussed here. 25 thing mammal fish shark whale orca dolphin Figure 5: A concept ....

[Article contains additional citation context not shown here]

Ardissono, L., and Goy, A. (2000b). Tailoring the Interaction with Users in Web Stores. User Modeling and User-Adapted Interaction 10(4): 251-303.

No context found.

L. Ardissono, A. Goy. Tailoring the interaction with users in web stores. User Modeling and User-Adapted Interaction, 10(4), 2000.

No context found.

Ardissono, L. and Goy, A. (2000): Tailoring the interaction with users in Web stores. User Modeling and User-Adapted Interaction, 10(4), 251--303.

No context found.

L. Ardissono and A. Goy. Tailoring the interaction with users in Web stores. User Modeling and User-Adapted Interaction, 10(4):251--303, 2000.

....the products and services they search for. This raises the following issue: how can a Web based system support these customers in finding the goods best fulfilling their needs To face this issue, several Web based systems offer one to one recommendation of items, given the customer s preferences [1, 2, 7, 10, 13, 15]. However, the recommendation techniques developed so far do not support the configuration of items, that is essential to comply with the customer s requirements when purchasing complex products, or registering for services. At the current stage, this type of activity can be performed by using ....

L. Ardissono and A. Goy, 'Tailoring the interaction with users in Web stores', User Modeling and User-Adapted Interaction, 10(4), 21-303, (200).

....when no information about her his watching behavior is available, except for her his explicit preferences. This module exploits the user s personal data and declared preferences to classify her him into a set of relevant user typologies and predict interests and watching behavior accordingly [1, 9]. The Dynamic UM Expert analyzes the user s watching behavior to identify her his preferences for program categories and update its local user model ( Dpref ) accordingly. This module also stores information about the most frequent keywords characterizing the programs watched by the user, in order ....

L. Ardissono and A. Goy. Tailoring the interaction with users in web stores. User Modeling and User-Adapted Interaction, 10(4):251-303, 2000.

....not as many helping texts than before. Besides the amount of information to be displayed, several further aspects of the interaction can be customised: For instance, the layout of The CAWICOMS Project the interface, or the type of questions asked for during the configuration of the product (see [1] for the definition of personalization strategies for a B2C application) In this description, we will focus on the last aspect, which strongly depends on the user s knowledgeability and interests and is critical to the usability of the configuration system. The basic idea is that the information ....

Ardissono L. and Goy A., Tailoring the Interaction With Users in Web stores. User Modeling and User-Adapted Interaction, Vol. 10(4), Kluwer, 2000, pp. 251-303.

.... viewpoint, we decided to build our stereotypes knowledge base starting from a lifestyles study, Sinottica, conducted by Eurisko data analyzers [4] However, the information regarding the lifestyles is not defined in a formalized way; thus, we chose our own representation format, based on [1]. We assumed a plausible correlation among homogeneous user groups and their preferences, then we structured the stereotypes in two main parts: i) a profile, containing the classification data of individuals belonging to the represented stereotype; ii) a prediction part, containing the typical ....

....TV watching frequency. The second one represents the situations where user preferences may occur during four time intervals in which the day is subdivided (morning, afternoon, evening, night) STRUCTURE OF THE STEREOTYPES Similar to the representation of stereotypical information adopted in SETA [1], we defined a family of stereotypes describing lifestyles. The features of a stereotype are represented as slots, according to the formalism introduced in Torasso and Console [9] Each slot includes an Importance facet, representing the impact of the feature on the overall description of the ....

Ardissono L., Goy A. (2000): Tailoring the Interaction With Users in Web stores. User Modeling and User-Adapted Interaction, 10(4), pp. 251-303, Kluwer Academic Publishers.

....the user. The user may also decide to skip the form; in this case we consider a default user. According to the conceptual decomposition of user modeling discussed above, we consider separate sets (families) of stereotypes for each one of the four dimensions that constitute the user model (see [2]) The stereotypes use the data provided by the user in the registration form as classificatory information and make predictions on different features of the user. While there is a partial overlap between the classificatory data used by different families of stereotypes, the predictions are not ....

L. Ardissono and A. Goy. Tailoring the interaction with users in web stores. User Modeling and User-Adapted Interaction, 10(4):251--303, 2000.

....strategies for selecting the information to be presented on the basis of the user s interests and familiarity with the products. Moreover, the system presents the available items for a product class (e.g. the available fax models) sorting them on the basis of the user s preferences [Ardissono and Goy, 2000] . During the interaction, the system monitors the user s selections to identify her needs for product functionalities and suggest potentially interesting product classes which the user has not visited. In this way, the assistance is extended to the search for alternative products. 4 Architecture ....

....they match the user s preferences. The Personalization Agent dynamically generates the code for the catalog pages by exploiting the information about the user provided by the UMC and a set of personalization rules for selecting information about products and type of description to be produced [Ardissono and Goy, 2000] . 5 Management of parallel user sessions The multi user access to the Web store is managed by performing the session tracking within the Session Manager (a Servlet) and forwarding the session specific messages to the agents of the architecture, so that they can process such messages and ....

[Article contains additional citation context not shown here]

L. Ardissono and A. Goy. Tailoring the interaction with users in web stores. User Modeling and User-Adapted Interaction, 10(4):251-- 303, 2000.

No context found.

L. Ardissono and A. Goy, `Tailoring the interaction with users in Web stores', User Modeling and User-Adapted Interaction, 10(4), 251--303, (2000).

No context found.

L. Ardissono and A. Goy. Tailoring the interaction with users in web stores. User Modeling and User-Adapted Interaction, 10(4):251--303, 2000.

No context found.

Ardissono L, Goy A, Tailoring the interaction with users in web stores, User Modelling and User Adapted Interaction, Vol. 10(4), 2000, pp. 251-303.

No context found.

Ardissono L., Goy A. (2000): Tailoring the Interaction With Users in Web stores. To appear on User Modeling and User-Adapted Interaction. Kluwer Academic Publishers.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC