127 research outputs found
Abstract Automated Construction of Relational Attributes ACORA: A Progress Report
No full textData mining research has not only development a large number of algorithms, but also enhanced our knowledge and understanding of their applicability and performance. However, the application of data mining technology in business environments is sdll no very common, despite the fact that organizations have access to large amounts of data and make deasions that could profit from data mining on a daily basis. One of the reasons is the mismatch between data representation for data storage and data analysis. Data are most commonly stored in multi-table relational databases whereas data mining methods require that the data be represented as a simple feature vector. This work presents a general framework for feature construction from multiple relational tables for data mining applications. The second part describes our prototype implementatio
Automated Construction of Relational Attributes ACORA: A Progress Report
Get PDFData mining research has not only development a large number of algorithms, but also
enhanced our knowledge and understanding of their applicability and performance.
However, the application of data mining technology in business environments is still no
very common, despite the fact that organizations have access to large amounts of data
and make decisions that could profit from data mining on a daily basis. One of the
reasons is the mismatch between data representation for data storage and data analysis.
Data are most commonly stored in multi-table relational databases whereas data mining
methods require that the data be represented as a simple feature vector. This work
presents a general framework for feature construction from multiple relational tables for
data mining applications. The second part describes our prototype implementation
ACORA (Automated Construction of Relational Features).Information Systems Working Papers Serie
Approaching the ILP 2005 Challenge: Class-Conditional Bayesian Propositionalization for Genetic Classification
No full textAbstract. This report presents a statistical propositionalisation approach to relational classification and probability estimation on the genetic ILP Challenge domain. The main difference between our and existing propositionalisation approaches is its ability to construct features from categorical attributes with many possible values and in particular the object identifiers. Our classification and ranking results on the genetic domain are promising but will require further evaluation in comparison with other relational models.
On cross-validation and stacking
No full textA number of times when using cross-validation (CV) while trying to do classification/probability estimation we have observed surprisingly low AUC's on real data with very few positive examples. AUC is the area under the ROC and measures the ranking ability and corresponds to the probability that a positive example receives a higher model score than a negative example. Intuition seems to suggest that no reasonable methodology should ever result in a model with an AUC significantly below 0.5. The focus of this paper is not on the estimator properties of CV (bias/variance/significance), but rather on the properties of the 'holdout' predictions based on which the CV performance of a model is calculated. We show that CV creates predictions that have an 'inverse' ranking with AUC well below 0.25 using features that were initially entirely unpredictive and models that can only perform monotonic transformations. In the extreme, combining CV with bagging (repeated averaging of out-of-sample predictions) generates 'holdout' predictions with perfectly opposite rankings on random data. While this would raise immediate suspicion upon inspection, we would like to caution the data mining community against using CV for stacking or in currently popular ensemble methods. They can reverse the predictions by assigning negative weights and produce in the end a model that appears to have close to perfect predictability while in reality the data was random.</jats:p
ACORA: Distribution-Based Aggregation for Relational Learning from Identifier Attributes
Get PDFFeature construction through aggregation plays an essential role in modeling relational
domains with one-to-many relationships between tables. One-to-many relationships
lead to bags (multisets) of related entities, from which predictive information
must be captured. This paper focuses on aggregation from categorical attributes
that can take many values (e.g., object identifiers). We present a novel aggregation
method as part of a relational learning system ACORA, that combines the use of
vector distance and meta-data about the class-conditional distributions of attribute
values. We provide a theoretical foundation for this approach deriving a "relational
fixed-effect" model within a Bayesian framework, and discuss the implications of
identifier aggregation on the expressive power of the induced model. One advantage
of using identifier attributes is the circumvention of limitations caused either by
missing/unobserved object properties or by independence assumptions. Finally, we
show empirically that the novel aggregators can generalize in the presence of identi-
fier (and other high-dimensional) attributes, and also explore the limitations of the
applicability of the methods.Information Systems Working Papers Serie
- …
