Abstract Sentiment analysis is an emerging technique that caters for semantic orientation and opinion mining. It is increasingly used to analyze online reviews and posts for identifying people's opinions and attitudes to products and events in order to improve business performance of companies and aid to make better organizing strategies of events. This paper presents an innovative approach to combining the outputs of sentiment classifiers under the framework of belief functions. It consists of the formulation of sentiment classifier outputs in the triplet evidence structure and the development of general formulas for combining triplet functions derived from sentiment classification results via three evidential combination rules along with comparative analyses. The empirical studies have been conducted on examining the effectiveness of our method for sentiment classification individually and in combination, and the results demonstrate that the best combined classifiers by our method outperforms the best individual classifiers over five review datasets.

In this work we investigate the use of parametric statistical methods for Anomaly Detection in time series data. The approach involves the use of simple and computationally efficient algorithms, the Cumulative Sum (CUSUM) and Exponentially Weighted Moving Average (EWMA), that have demonstrated an acceptable performance in detecting different shifts from the process mean. However, while the performance of these algorithms is found to be adequate in datasets where anomalies have a profound form, they produce many false positives when anomalies become more complex. To address this limitation, we propose a solution that has greater flexibility, in the form of a combined CUSUM-EWMA algorithm. Four different statistical methods are investigated and implemented, including the classic CUSUM and EWMA, and two variants of a combined CUSUM-EWMA algorithm. These algorithms have been evaluated on ten benchmark datasets. The F-Score for each of the algorithms has been used to measure their performance appropriately. The preliminary experimental results prove to be promising for the proposed method in detecting anomalies from time series data.

In this work we investigate the use of a fuzzy inspired approach for anomaly detection in different electromagnetic sequential time series datasets. The method proposed consists of simple component methods that are aggregated in a serialized way to achieve anomaly detection. Each of the component methods adds an element towards anomaly detection, i.e. a smoothing filter removes any unwanted noise, an automated peak finding with Fast Fourier Transformation and correlation, reduces the dimensionality of the signal, a fuzzy inference system encodes the signal before the final comparison and its respective output. This method is evaluated on 6 benchmark datasets with promising results in terms of F-measure accuracy. The method is also evaluated over real datasets gathered from the SWARM satellites for the detection of possible anomalies in relation to seismic events. The preliminary experimental results also prove to be promising for the proposed method for the detection of anomalies in electromagnetic sequential time series datasets.

Data fusion is currently used extensively in information retrieval for various tasks. It has proved to be a useful technology because it is able to improve retrieval performance frequently. However, in almost all prior research in data fusion, static search environments have been used, and dynamic search environments have generally not been considered. In this article, we investigate adaptive data fusion methods that can change their behavior when the search environment changes. Three adaptive data fusion methods are proposed and investigated. To test these proposed methods properly, we generate a benchmark from a historic T ext RE trieval Conference data set. Experiments with the benchmark show that 2 of the proposed methods are good and may potentially be used in practice.

One of the most popular techniques of generating classifier ensembles is known as stacking which is based on a meta-learning approach. In this paper, we introduce an alternative method to stacking which is based on cluster analysis. Similar to stacking, instances from a validation set are initially classified by all base classifiers. The output of each classifier is subsequently considered as a new attribute of the instance. Following this, a validation set is divided into clusters according to the new attributes and a small subset of the original attributes of the instances. For each cluster, we find its centroid and calculate its class label. The collection of centroids is considered as a meta-classifier. Experimental results show that the new method outperformed all benchmark methods, namely Majority Voting, Stacking J48, Stacking LR, AdaBoost J48, and Random Forest, in 12 out of 22 data sets. The proposed method has two advantageous properties: it is very robust to relatively small training sets and it can be applied in semi-supervised learning problems. We provide a theoretical investigation regarding the proposed method. This demonstrates that for the method to be successful, the base classifiers applied in the ensemble should have greater than 50% accuracy levels.

The Latent Dirichlet Allocation model is an unsupervised generative model that is widely used for topic modelling in text. We propose to add supervision to the model in the form of domain knowledge to direct the focus of topics to more relevant aspects than the topics produced by standard LDA. Experimental results demonstrate the effectiveness of our method. We also propose a novel Twofold-LDA model to improve the current output of LDA in order to visualize results in graphical form, which can ultimately be used by potential customers. Experiments show the benefit of this new output, with the ability to produce topics focused on our desired aspects in a user friendly chart.

Linear-time computational techniques based on the structure of an evidence space have been developed for combining multiple pieces of evidence using Dempster's rule (orthogonal sum), which is available on a number of contending hypotheses. They offer a means of making the computation-intensive calculations involved more efficient in certain circumstances. Unfortunately, they restrict the orthogonal sum of evidential functions to the dichotomous structure that applies only to elements and their complements. In this paper, we present a novel evidence structure in terms of a triplet and a set of algorithms for evidential reasoning. The merit of this structure is that it divides a set of evidence into three subsets, distinguishing the trivial evidential elements from the important ones—focusing particularly on some elements of an evidence space. It avoids the deficits of the dichotomous structure in representing the preference of evidence and estimating the basic probability assignment of evidence. We have established a formalism for this structure and the general formulae for combining pieces of evidence in the form of the triplet, which have been theoretically and empirically justified. © 2008 Wiley Periodicals, Inc.

Experience shows that different text classification methods can give different results. We look here at a way of combining the results of two or more different classification methods using an evidential approach. The specific methods we have been experimenting with in our group include the support vector machine, kNN (nearest neighbors), kNN model-based approach (kNNM), and Rocchio methods, but the analysis and methods apply to any methods. We review these learning methods briefly, and then we describe our method for combining the classifiers. In a previous study, we suggested that the combination could be done using evidential operations and that using only two focal points in the mass functions gives good results. However, there are conditions under which we should choose to use more focal points. We assess some aspects of this choice from an reasoning perspective and suggest a refinement of the approach.