Rafael Giusti

Spoken Digit Recognition in Portuguese Using Line Spectral Frequencies

Recognition of isolated spoken digits is the core procedure for a large and important number of applications mainly in telephone based services, such as dialing, airline reservation, bank transaction and price quotation, only using speech. Spoken digit recognition is generally a challenging task since the signals last for short period of time and often some digits are acoustically very similar to each other. The objective of this paper is to investigate the use of machine learning algorithms for digit recognition. We focus on the recognition of digits spoken in Portuguese. However, we note that our techniques are applicable to any language. We believe that the most important task for successfully recognizing spoken digits is the attribute extraction. Audio data is composed by a huge amount of very weak features, and most machine learning algorithms will not be able to build accurate classifiers. We show that Line Spectral Frequencies (LSF) provides a set of highly predictive coefficients for digit recognition. The results are superior than those obtained with state-of-the-art methods using Mel-Frequency Cepstrum Coefficients (MFCC) for digit recognition. In particular, we show that the choice of the right attribute extraction method is more important than the specific classification paradigm, and that the right combination of classifier and attributes can provide almost perfect accuracy.

A Multi-Objective Evolutionary Algorithm to Build Knowledge Classification Rules with Specific Properties

This work proposes the use of evolutionary algorithms to build individual knowledge rules with specific properties that are usually neglected when conducted by traditional supervised learning methods. The proposed evolutionary algorithm uses a rank-based, multi-objective fitness function that enables the arrangement of any set of measures. Experimental results that show the suitability of our proposal are also presented.

Speeding up similarity search under dynamic time warping by pruning unpromising alignments

Similarity search is the core procedure for several time series mining tasks. While different distance measures can be used for this purpose, there is clear evidence that the Dynamic Time Warping (DTW) is the most suitable distance function for a wide range of application domains. Despite its quadratic complexity, research efforts have proposed a significant number of pruning methods to speed up the similarity search under DTW. However, the search may still take a considerable amount of time depending on the parameters of the search, such as the length of the query and the warping window width. The main reason is that the current techniques for speeding up the similarity search focus on avoiding the costly distance calculation between as many pairs of time series as possible. Nevertheless, the few pairs of subsequences that were not discarded by the pruning techniques can represent a significant part of the entire search time. In this work, we adapt a recently proposed algorithm to improve the internal efficiency of the DTW calculation. Our method can speed up the UCR suite, considered the current fastest tool for similarity search under DTW. More important, the longer the time needed for the search, the higher the speedup ratio achieved by our method. We demonstrate that our method performs similarly to UCR suite for small queries and narrow warping constraints. However, it performs up to five times faster for long queries and large warping windows.

Time Series Classification with Representation Ensembles

Time series has attracted much attention in recent years, with thousands of methods for diverse tasks such as classification, clustering, prediction, and anomaly detection. Among all these tasks, classification is likely the most prominent task, accounting for most of the applications and attention from the research community. However, in spite of the huge number of methods available, there is a significant body of empirical evidence indicating that the 1-nearest neighbor algorithm (\(1\)-NN) in the time domain is “extremely difficult to beat”. In this paper, we evaluate the use of different data representations in time series classification. Our work is motivated by methods used in related areas such as signal processing and music retrieval. In these areas, a change of representation frequently reveals features that are not apparent in the original data representation. Our approach consists of using different representations such as frequency, wavelets, and autocorrelation to transform the time series into alternative decision spaces. A classifier is then used to provide a classification for each test time series in the alternative domain. We investigate how features provided in different domains can help in time series classification. We also experiment with different ensembles to investigate if the data representations are a good source of diversity for time series classification. Our extensive experimental evaluation approaches the issue of combining sets of representations and ensemble strategies, resulting in over 300 ensemble configurations.

Discovering Knowledge Rules with Multi-Objective Evolutionary Computing

Most Machine Learning systems target into inducing classifiers with optimal coverage and precision measures. Although this constitutes a good approach for prediction, it might not provide good results when the user is more interested in description. In this case, the induced models should present other properties such as novelty, interestingness and so forth. In this paper we present a research work based in Multi-Objective Evolutionary Computing to construct individual knowledge rules targeting arbitrary user-defined criteria via objective quality measures such as precision, support, novelty etc. This paper also presents a comparison among multi-objective and ranking composition techniques. It is shown that multi-objective-based methods attain better results than ranking-based methods, both in terms of solution dominance and diversity of solutions in the Pareto front.

Asfault: A low-cost system to evaluate pavement conditions in real-time using smartphones and machine learning

Abstract Modern smartphones have a large variety of built-in sensors that can measure different information about users and the environment around them. Given the increasing popularity of these devices, their high processing power, and the ability to transfer data over wireless networks, different smartphone-based applications have emerged in the last years to solve old problems with new approaches more efficiently and cheaply. One example is the assessment and monitoring of asphalt quality. This task has been done manually by experts since the 1930s, and with the help of expensive equipment since the 1960s. Currently, we are experiencing the emergence of next-generation tools to perform this monitoring with smartphones, significantly reducing costs, time, and effort of experts. However, there is a trade-off between the costs and precision of smartphone sensors, requiring the use of sophisticated software solutions. In this paper, we propose Asfault, a low-cost system to evaluate and monitor road pavement conditions in real-time using smartphone sensors and machine learning algorithms. The system is composed of an Android application responsible for doing automatic evaluations and a web application that aims to show the evaluations in an informative way. We propose to employ accelerometer sensors to measure the vehicle vibration while driving and use this data to evaluate the pavement conditions. Asfault achieves a classification performance superior to 90% in a 5-class problem considering the following road qualities: Good, Average, Fair, and Poor, as well the occurrence of obstacles in the road. Our system is publicly available for use and could be useful for practitioners responsible for urban and highway maintenance, as well for regular drivers in the planning of better routes based on the pavement quality and comfort of the travel.

Improved Time Series Classification with Representation Diversity and SVM

Time series classification is an important task in data mining that has been traditionally addressed with the use of similarity-based classifiers. The 1-NN DTW is typically considered the most accurate model for temporal data. Nevertheless, some authors have recently proposed ingenious alternatives to the 1-NN DTW by using diversity of time series representation or by using DTW for feature extraction. In this paper, we explore diversity of time series representations and distance functions to obtain distance features, which in turn are used to train an SVM model. We argue that the scientific community has largely neglected a vast body of unconventional distance functions, and we present empirical evidence that distance features are better than the 1-NN DTW with respect to classification accuracy.