Welcome To INCOGNITOMATCH
Explore
OVERVIEW
Matching schemata is the first step in this process. Schema matching aims to find correspondences between concepts describing the meaning of data sources. Human-in-the-loop data integration, and specifically in matching, has been recently challenged by the big data era. With richer data source variety, bigger volumes, and faster velocity, accessing and making use of humans’ input has become much harder.
Acknowledging cognitive awareness in human matching, InCognitoMatch (Introducing Cognitive Biases to Crowdsourced Matching) is a handy tool to validate, annotate, and correct correspondences using the crowd.
THE TASK
The task begins with providing personal information (e.g., ID and nickname) and a (short) intelligence test.
Subsequently, a subset of matching correspondences will be introduced. At the end, you will be able to analyze your performance using a dedicated dashboard.
In the main part of the session, several element pairs will be introduced. Each pair consists of two terms, one from a candidate schema and one from target schema. To accompany each pair, additional properties will be introduced including term names and types, instance examples, schemata hierarchy, algorithmic matching result and a majority decision of the other participants.
Your task is to provide feedback, stating your confidence level (on a [0,100] scale) about whether a pair is a match (the two terms correspond) or no match.
Confidence level higher than 50 indicates preference for match while value lower than 50 indicates preference for no match. Please note that you will not be able to choose confidence level of 50.
EXAMPLE
The figure above illustrates the matching task using a matching question regarding the pair contactName (Term A) and contactName (Term B).
The two sides of screen show information about the two terms. Each term name appears clearly on top and marked in red in the center of the screen.
Following the name, the data types and instance examples are introduced. In this example, the types of both contactName (Term A) and contactName (Term B) are “str”. An example instance of contactName (Term A) is "Dr. Gilboa Barak" and contactName (Term B) instance is "Ms Einat Hadar".
The center part of the screen displays the schemata hierarchy. The sign indicates the term's level within the hierarchy. For example, the term contactName (Term A) is placed at the third level.
The color legend is as follows:
Green - Parents of the term.
Red - The term.
Blue - Siblings of the term.
The direct parent of contactName (Term A) is Contact and a sibling term is, for example, telephone.
As demonstrated below, the algorithm matching result is 100% and the majority decision of the crowd is 66.67%.
Finally, the decision should be made by selecting the confidence level and then press on the SUBMIT button.
TASK SETTINGS
By proceeding to the task, you approve having your personal information and task results stored in the system. This data will be used for academic research only.
Click on the TASK button.