Paula Branco

A Survey of Predictive Modeling on Imbalanced Domains

Many real-world data-mining applications involve obtaining predictive models using datasets with strongly imbalanced distributions of the target variable. Frequently, the least-common values of this target variable are associated with events that are highly relevant for end users (e.g., fraud detection, unusual returns on stock markets, anticipation of catastrophes, etc.). Moreover, the events may have different costs and benefits, which, when associated with the rarity of some of them on the available training data, creates serious problems to predictive modeling techniques. This article presents a survey of existing techniques for handling these important applications of predictive analytics. Although most of the existing work addresses classification tasks (nominal target variables), we also describe methods designed to handle similar problems within regression tasks (numeric target variables). In this survey, we discuss the main challenges raised by imbalanced domains, propose a definition of the problem, describe the main approaches to these tasks, propose a taxonomy of the methods, summarize the conclusions of existing comparative studies as well as some theoretical analyses of some methods, and refer to some related problems within predictive modeling.

SMOTE for Regression

Several real world prediction problems involve forecasting rare values of a target variable. When this variable is nominal we have a problem of class imbalance that was already studied thoroughly within machine learning. For regression tasks, where the target variable is continuous, few works exist addressing this type of problem. Still, important application areas involve forecasting rare extreme values of a continuous target variable. This paper describes a contribution to this type of tasks. Namely, we propose to address such tasks by sampling approaches. These approaches change the distribution of the given training data set to decrease the problem of imbalance between the rare target cases and the most frequent ones. We present a modification of the well-known Smote algorithm that allows its use on these regression tasks. In an extensive set of experiments we provide empirical evidence for the superiority of our proposals for these particular regression tasks. The proposed SmoteR method can be used with any existing regression algorithm turning it into a general tool for addressing problems of forecasting rare extreme values of a continuous target variable.

Resampling strategies for regression

Several real world prediction problems involve forecasting rare values of a target variable. When this variable is nominal, we have a problem of class imbalance that was thoroughly studied within machine learning. For regression tasks, where the target variable is continuous, few works exist addressing this type of problem. Still, important applications involve forecasting rare extreme values of a continuous target variable. This paper describes a contribution to this type of tasks. Namely, we propose to address such tasks by resampling approaches that change the distribution of the given data set to decrease the problem of imbalance between the rare target cases and the most frequent ones. We present two modifications of well-known resampling strategies for classification tasks: the under-sampling and the synthetic minority over-sampling technique SMOTE methods. These modifications allow the use of these strategies on regression tasks where the goal is to forecast rare extreme values of the target variable. In an extensive set of experiments, we provide empirical evidence for the superiority of our proposals for these particular regression tasks. The proposed resampling methods can be used with any existing regression algorithm, which means that they are general tools for addressing problems of forecasting rare extreme values of a continuous target variable.

A Framework for Recommendation of Highly Popular News Lacking Social Feedback

Social media is rapidly becoming the main source of news consumption for users, raising significant challenges to news aggregation and recommendation tasks. One of these challenges concerns the recommendation of very recent news. To tackle this problem, approaches to the prediction of news popularity have been proposed. In this paper, we study the task of predicting news popularity upon their publication, when social feedback is unavailable or scarce, and to use such predictions to produce news rankings. Unlike previous work, we focus on accurately predicting highly popular news. Such cases are rare, causing known issues for standard prediction models and evaluation metrics. To overcome such issues we propose the use of resampling strategies to bias learners towards these rare cases of highly popular news, and a utility-based framework for evaluating their performance. An experimental evaluation is performed using real-world data to test our proposal in distinct scenarios. Results show that our proposed approaches improve the ability of predicting and recommending highly popular news upon publication, in comparison to previous work.

Resampling Strategies for Imbalanced Time Series

Time series forecasting is a challenging task, where the non-stationary characteristics of the data portrays a hard setting for predictive tasks. A common issue is the imbalanced distribution of the target variable, where some intervals are very important to the user but severely underrepresented. Standard regression tools focus on the average behaviour of the data. However, the objective is the opposite in many forecasting tasks involving time series: predicting rare values. A common solution to forecasting tasks with imbalanced data is the use of resampling strategies, which operate on the learning data by changing its distribution in favor of a given bias. The objective of this paper is to provide solutions capable of significantly improving the predictive accuracy of rare cases in forecasting tasks using imbalanced time series data. We extend the application of resampling strategies to the time series context and introduce the concept of temporal and relevance bias in the case selection process of such strategies, presenting new proposals. We evaluate the results of standard regression tools and the use of resampling strategies, with and without bias over 24 time series data sets from 6 different sources. Results show a significant increase in predictive accuracy of rare cases associated with the use of resampling strategies, and the use of biased strategies further increases accuracy over the non-biased strategies.

Exploring Resampling with Neighborhood Bias on Imbalanced Regression Problems

Imbalanced domains are an important problem that arises in predictive tasks causing a loss in the performance of the most relevant cases for the user. This problem has been intensively studied for classification problems. Recently it was recognized that imbalanced domains occur in several other contexts and for a diversity of types of tasks. This paper focus on imbalanced regression tasks. Resampling strategies are among the most successful approaches to imbalanced domains. In this work we propose variants of existing resampling strategies that are able to take into account the information regarding the neighborhood of the examples. Instead of performing sampling uniformly, our proposals bias the strategies for reinforcing some regions of the data sets. In an extensive set of experiments we provide evidence of the advantage of introducing a neighborhood bias in the resampling strategies.

Relevance-Based Evaluation Metrics for Multi-class Imbalanced Domains

The class imbalance problem is a key issue that has received much attention. This attention has been mostly focused on two-classes problems. Fewer solutions exist for the multi-classes imbalance problem. From an evaluation point of view, the class imbalance problem is challenging because a non-uniform importance is assigned to the classes. In this paper, we propose a relevance-based evaluation framework that incorporates user preferences by allowing the assignment of differentiated importance values to each class. The presented solution is able to overcome difficulties detected in existing measures and increases discrimination capability. The proposed framework requires the assignment of a relevance score to the problem classes. To deal with cases where the user is not able to specify each class relevance, we describe three mechanisms to incorporate the existing domain knowledge into the relevance framework. These mechanisms differ in the amount of information available and assumptions made regarding the domain. They also allow the use of our framework in common settings of multi-class imbalanced problems with different levels of information available.

Resampling strategies for imbalanced time series forecasting

Time series forecasting is a challenging task, where the non-stationary characteristics of data portray a hard setting for predictive tasks. A common issue is the imbalanced distribution of the target variable, where some values are very important to the user but severely under-represented. Standard prediction tools focus on the average behaviour of the data. However, the objective is the opposite in many forecasting tasks involving time series: predicting rare values. A common solution to forecasting tasks with imbalanced data is the use of resampling strategies, which operate on the learning data by changing its distribution in favour of a given bias. The objective of this paper is to provide solutions capable of significantly improving the predictive accuracy on rare cases in forecasting tasks using imbalanced time series data. We extend the application of resampling strategies to the time series context and introduce the concept of temporal and relevance bias in the case selection process of such strategies, presenting new proposals. We evaluate the results of standard forecasting tools and the use of resampling strategies, with and without bias over 24 time series data sets from six different sources. Results show a significant increase in predictive accuracy on rare cases associated with using resampling strategies, and the use of biased strategies further increases accuracy over non-biased strategies.

Crime Prediction Using Regression and Resources Optimization

Violent crime is a well known social problem affecting both the quality of life and the economical development of a society. Its prediction is therefore an important asset for law enforcement agencies, since due to budget constraints, the optimization of resources is of extreme importance. In this work, we tackle both aspects: prediction and optimization.