Advances in Data Analysis and Classification最新文献

Mixture models provide a useful tool to account for unobserved heterogeneity and are at the basis of many model-based clustering methods. To gain additional flexibility, some model parameters can be expressed as functions of concomitant covariates. In this Paper, a semiparametric finite mixture of regression models is defined, with concomitant information assumed to influence both the component weights and the conditional means. In particular, linear predictors are replaced with smooth functions of the covariate considered by resorting to cubic splines. An estimation procedure within the Bayesian paradigm is suggested, where smoothness of the covariate effects is controlled by suitable choices for the prior distributions of the spline coefficients. A data augmentation scheme based on difference random utility models is exploited to describe the mixture weights as functions of the covariate. The performance of the proposed methodology is investigated via simulation experiments and two real-world datasets, one about baseball salaries and the other concerning nitrogen oxide in engine exhaust.

Sentiment analysis or opinion mining refers to the discovery of sentiment information within textual documents, tweets, or review posts. This field has emerged with the social media outgrowth which becomes of great interest for several applications such as marketing, tourism, and business. In this work, we approach Twitter sentiment analysis through a novel framework that addresses simultaneously the problems of text representation such as sparseness and high-dimensionality. We propose an information retrieval probabilistic model based on a new distribution namely the Smoothed Scaled Dirichlet distribution. We present a likelihood learning method for estimating the parameters of the distribution and we propose a feature generation from the information retrieval system. We apply the proposed approach Smoothed Scaled Relevance Model on four Twitter sentiment datasets: STD, STS-Gold, SemEval14, and SentiStrength. We evaluate the performance of the offered solution with a comparison against the baseline models and the related-works.

The number of daily credit card transactions is inexorably growing: the e-commerce market expansion and the recent constraints for the Covid-19 pandemic have significantly increased the use of electronic payments. The ability to precisely detect fraudulent transactions is increasingly important, and machine learning models are now a key component of the detection process. Standard machine learning techniques are widely employed, but inadequate for the evolving nature of customers behavior entailing continuous changes in the underlying data distribution. his problem is often tackled by discarding past knowledge, despite its potential relevance in the case of recurrent concepts. Appropriate exploitation of historical knowledge is necessary: we propose a learning strategy that relies on diversity-based ensemble learning and allows to preserve past concepts and reuse them for a faster adaptation to changes. In our experiments, we adopt several state-of-the-art diversity measures and we perform comparisons with various other learning approaches. We assess the effectiveness of our proposed learning strategy on extracts of two real datasets from two European countries, containing more than 30 M and 50 M transactions, provided by our industrial partner, Worldline, a leading company in the field.

Clustering mixed-type data, that is, observation by variable data that consist of both continuous and categorical variables poses novel challenges. Foremost among these challenges is the choice of the most appropriate clustering method for the data. This paper presents a benchmarking study comparing eight distance-based partitioning methods for mixed-type data in terms of cluster recovery performance. A series of simulations carried out by a full factorial design are presented that examined the effect of a variety of factors on cluster recovery. The amount of cluster overlap, the percentage of categorical variables in the data set, the number of clusters and the number of observations had the largest effects on cluster recovery and in most of the tested scenarios. KAMILA, K-Prototypes and sequential Factor Analysis and K-Means clustering typically performed better than other methods. The study can be a useful reference for practitioners in the choice of the most appropriate method.

Symbolic data analysis (SDA) is an emerging area of statistics concerned with understanding and modelling data that takes distributional form (i.e. symbols), such as random lists, intervals and histograms. It was developed under the premise that the statistical unit of interest is the symbol, and that inference is required at this level. Here we consider a different perspective, which opens a new research direction in the field of SDA. We assume that, as with a standard statistical analysis, inference is required at the level of individual-level data. However, the individual-level data are unobserved, and are aggregated into observed symbols—group-based distributional-valued summaries—prior to the analysis. We introduce a novel general method for constructing likelihood functions for symbolic data based on a desired probability model for the underlying measurement-level data, while only observing the distributional summaries. This approach opens the door for new classes of symbol design and construction, in addition to developing SDA as a viable tool to enable and improve upon classical data analyses, particularly for very large and complex datasets. We illustrate this new direction for SDA research through several real and simulated data analyses, including a study of novel classes of multivariate symbol construction techniques.

Flexible modelling of interval-valued data is of great practical importance with the development of advanced technologies in current data collection processes. This paper proposes a new robust regression model for interval-valued data based on midpoints and log-ranges of the dependent intervals, and obtains the parameter estimators using Huber loss function to deal with possible outliers in a data set. Besides, the use of logarithm transformation avoids the non-negativity constraints for the traditional modelling of ranges, which is beneficial to the flexible use of various regression methods. We conduct extensive Monte Carlo simulation experiments to compare the finite-sample performance of our model with that of the existing regression methods for interval-valued data. Results indicate that the proposed method shows competitive performance, especially in the data set with the existence of outliers and the scenarios where both midpoints and ranges of independent variables are related to those of the dependent one. Moreover, two empirical interval-valued data sets are applied to illustrate the effectiveness of our method.

The k-Means algorithm is one of the most popular choices for clustering data but is well-known to be sensitive to the initialization process. There is a substantial number of methods that aim at finding optimal initial seeds for k-Means, though none of them is universally valid. This paper presents an extension to longitudinal data of one of such methods, the BRIk algorithm, that relies on clustering a set of centroids derived from bootstrap replicates of the data and on the use of the versatile Modified Band Depth. In our approach we improve the BRIk method by adding a step where we fit appropriate B-splines to our observations and a resampling process that allows computational feasibility and handling issues such as noise or missing data. We have derived two techniques for providing suitable initial seeds, each of them stressing respectively the multivariate or the functional nature of the data. Our results with simulated and real data sets indicate that our Functional Data Approach to the BRIK method (FABRIk) and our Functional Data Extension of the BRIK method (FDEBRIk) are more effective than previous proposals at providing seeds to initialize k-Means in terms of clustering recovery.

We consider nonparametric prediction with multiple covariates, in particular categorical or functional predictors, or a mixture of both. The method proposed bases on an extension of the Nadaraya-Watson estimator where a kernel function is applied on a linear combination of distance measures each calculated on single covariates, with weights being estimated from the training data. The dependent variable can be categorical (binary or multi-class) or continuous, thus we consider both classification and regression problems. The methodology presented is illustrated and evaluated on artificial and real world data. Particularly it is observed that prediction accuracy can be increased, and irrelevant, noise variables can be identified/removed by ‘downgrading’ the corresponding distance measures in a completely data-driven way.

Multiple imputation (MI) is a popular method for dealing with missing values. However, the suitable way for applying clustering after MI remains unclear: how to pool partitions? How to assess the clustering instability when data are incomplete? By answering both questions, this paper proposed a complete view of clustering with missing data using MI. The problem of partitions pooling is here addressed using consensus clustering while, based on the bootstrap theory, we explain how to assess the instability related to observed and missing data. The new rules for pooling partitions and instability assessment are theoretically argued and extensively studied by simulation. Partitions pooling improves accuracy, while measuring instability with missing data enlarges the data analysis possibilities: it allows assessment of the dependence of the clustering to the imputation model, as well as a convenient way for choosing the number of clusters when data are incomplete, as illustrated on a real data set.