Drift mining in data: A framework for addressing drift in classification

A novel statistical methodology for analysing population drift in classification is introduced. Drift denotes changes in the joint distribution of explanatory variables and class labels over time. It entails the deterioration of a classifier's performance and requires the optimal decision boundary to be adapted after some time. However, in the presence of verification latency a re-estimation of the classification model is impossible, since in such a situation only recent unlabelled data are available, and the true corresponding labels only become known after some lapse in time. For this reason a novel drift mining methodology is presented which aims at detecting changes over time. It allows us either to understand evolution in the data from an ex-post perspective or, ex-ante, to anticipate changes in the joint distribution. The proposed drift mining technique assumes that the class priors change by a certain factor from one time point to the next, and that the conditional distributions do not change within this time period. Thus, the conditional distributions can be estimated at a time where recent labelled data are available. In subsequent periods the unconditional distribution can be expressed as a mixture of the conditional distributions, where the mixing proportions are equal to the class priors. However, as the unconditional distributions can also be estimated from new unlabelled data, they can then be compared to the mixture representation by means of least-squares criteria. This allows for easy and fast estimation of the changes in class prior values in the presence of verification latency. The usefulness of this drift mining approach is demonstrated using a real-world dataset from the area of credit scoring. Highlights? A novel statistical methodology for analysing population drift in classification is presented. ? Drift mining that aims at modelling changes in distributions over time is introduced. ? A model on global drift that addresses a change in class prior is introduced.

[1]  Ludmila I. Kuncheva,et al.  Classifier Ensembles for Detecting Concept Change in Streaming Data: Overview and Perspectives , 2008 .

[2]  Niall M. Adams,et al.  The impact of changing populations on classifier performance , 1999, KDD '99.

[3]  Peter A. Flach,et al.  A Response to Webb and Ting’s On the Application of ROC Analysis to Predict Classification Performance Under Varying Class Distributions , 2005, Machine Learning.

[4]  D. Hand,et al.  Local Versus Global Models for Classification Problems , 2003 .

[5]  Thomas M. Cover,et al.  Elements of Information Theory , 2005 .

[6]  Richard Granger,et al.  Beyond Incremental Processing: Tracking Concept Drift , 1986, AAAI.

[7]  Jie Zhou,et al.  Transfer estimation of evolving class priors in data stream classification , 2010, Pattern Recognit..

[8]  Michaela M. Black,et al.  The Impact of Latency on Online Classification Learning with Concept Drift , 2010, KSEM.

[9]  Joachim M. Buhmann,et al.  Agnostic Domain Adaptation , 2011, DAGM-Symposium.

[10]  Christoforos Anagnostopoulos,et al.  Temporally-Adaptive Linear Classification for Handling Population Drift in Credit Scoring , 2010, COMPSTAT.

[11]  Göran Kauermann,et al.  Density estimation and comparison with a penalized mixture approach , 2012, Comput. Stat..

[12]  George Forman,et al.  Quantifying counts and costs via classification , 2008, Data Mining and Knowledge Discovery.

[13]  M. Durbán,et al.  Generalized linear array models with applications to multidimensional smoothing , 2006 .

[14]  Geoffrey I. Webb,et al.  On the Application of ROC Analysis to Predict Classification Performance Under Varying Class Distributions , 2005, Machine Learning.

[15]  Koby Crammer,et al.  Analysis of Representations for Domain Adaptation , 2006, NIPS.

[16]  Georg Krempl,et al.  Classification in Presence of Drift and Latency , 2011, 2011 IEEE 11th International Conference on Data Mining Workshops.

[17]  Rudolf Kruse,et al.  An Algorithm for Anticipating Future Decision Trees from Concept-Drifting Data , 2008, SGAI Conf..

[18]  Amos Storkey,et al.  When Training and Test Sets are Different: Characterising Learning Transfer , 2013 .

[19]  Francesco Tortorella.  An Optimal Reject Rule for Binary Classifiers , 2000, SSPR/SPR.

[20]  Qiang Yang,et al.  Boosting for transfer learning , 2007, ICML '07.

[21]  Jesús Cid-Sueiro,et al.  Class and subclass probability re-estimation to adapt a classifier in the presence of concept drift , 2011, Neurocomputing.

[22]  Alexander Zien,et al.  Semi-Supervised Learning , 2006 .

[23]  Marco Saerens,et al.  Adjusting the Outputs of a Classifier to New a Priori Probabilities: A Simple Procedure , 2002, Neural Computation.

[24]  Michael A. Stephens,et al.  Asymptotic Results for Goodness-of-Fit Statistics with Unknown Parameters , 1976 .

[25]  Žliobait . e,et al.  Learning under Concept Drift: an Overview , 2010 .

[26]  Padraig Cunningham,et al.  A case-based technique for tracking concept drift in spam filtering , 2004, Knowl. Based Syst..

[27]  Indre Zliobaite,et al.  Learning under Concept Drift: an Overview , 2010, ArXiv.

[28]  Alexey Tsymbal,et al.  The problem of concept drift: definitions and related work , 2004 .

[29]  David J. Hand,et al.  Measuring classifier performance: a coherent alternative to the area under the ROC curve , 2009, Machine Learning.

[30]  Georg Krempl,et al.  The Algorithm APT to Classify in Concurrence of Latency and Drift , 2011, IDA.

[31]  Myra Spiliopoulou,et al.  On exploiting the power of time in data mining , 2008, SKDD.