Last Wednesday, Dr. Cobb from the department of Computer Science and Statistics gave us a seminar about the conflation of vector mapping data.

First of all, Dr. Cobb introduced the concept of the conflation. She generalized the conflation as the following three phases:

• Feature Matching.
• Feature Deconfliction.
• Positional Alignment.

Afterwoards, She discussed the three phases in details.

The feature matching is accomplished by three methods:

• Pre-processing step to unify logical feature segment during conflation.
• Hierarchical rule based on matching criteria.
• User interaction.

Dr. Cobb emphasized that we need deconflict attribute set and values during the process of feature deconfliction. The attribute set varies based on the product.

Dr. Cobb then discussed the problem of the attribute value deconfliction. She discussed some important factors in the process of attribute value deconfliction. They are:

• Rule-based processing invovling information from data quality table and linear information.
• Rule-based processing alignmented with capabilities for reasoning under uncertainty.

Finally, Dr. Cobb discussed the potential problem of the Conflation of mapping data, such as missing data and lacking reliability of information. However, this may be adjusted by user interaction. She also discussed the problem of differnet scale mapping.