This innovative practice full paper presents an inter-disciplinary Project-Based Learning (IPBL) approach that relies on a challenge as a key part of the learning process. A project is proposed to the students from the perspective of creating a direct association between learning and doing. In this paper, we present a variation of PBL adding the interdisciplinary factor: we united students from different areas (digital games and music production) under the common objective of producing a complete sounded digital game. An additional difficulty was overcome since all classes took place remotely due to the Covid-19 pandemic. To accomplish this challenge, a remote structure of classes was created to facilitate interaction between students as much as possible. During one semester, 2 groups of undergraduates from these two areas studied different modules organized in parallel under the general objective of an integrative project. The integrated project spanned 17 weeks, with classes structured to accommodate the evolving needs of project development. Initially, the first three weeks operated independently, allowing for focused exploration within each module. Subsequently, collaborative efforts were emphasized during weeks 4, 6, 9, 10, 12, and 15, fostering synergy and collective progress across both domains. Game students were given the mission of producing a game from scratch. To music production students the mission was making a full package of sound and music to each respective game. Facing a common schedule, all students had periodic meetings to align production and deadlines. In the end 10 games were produced and sounded. Initial hypotheses were raised about what we expected because of this experience. To assist in the analysis of the results, all the students involved responded to surveys about what they experienced. The successes and the need for improvements observed are detailed in this paper based on the application of the proposed methodology during the period in question.

STEM (Science, Technology, Engineering, and Math) courses are male-dominated spaces where women are the minority most of the time. Even now, when women are the majority in universities, the computer science and engineering degrees are still uneven when it comes to gender. There have been interesting initiatives in the past Frontiers In Education (FIE) conferences about several of the issues related to bringing more girls into the computer science and engineering fields of study. Some of these works typically address one or more challenges associated with increasing the number of women in computer science and engineering by using different approaches and targeting different age groups. The majority of the studied works focus on raising awareness through the development of boot camps, workshops, programs and other initiatives. Other works investigate the attitude towards computing careers and the lack of interest in these areas by female students, regardless of age. There are also works that are focused on female student's retention and success in their academic degrees. All of the reviewed works have one major goal in common: bringing more girls into STEM fields, such as engineering and computer science. In order to achieve this goal, the existing initiatives employ disciplinary and interdisciplinary approaches. Although the majority of the work's main technical focus is on teaching programming, this is achieved by using a variety of approaches and resources. For example, there are works aimed at developing digital games, approaches that use robotics with Arduino and also approaches that integrate computer programming with musical instruments, among others. We did not find a study that aggregated and analyzed these initiatives. The main goal of this paper is to analyze past FIE editions to identify and aggregate studies related to attracting girls to STEM areas. The method was based on systematic review procedures considering FIE 2010 thru FIE 2022. We analyze and evaluate the existing works based on their objectives, learning topics, target group information, and availability of supplementary materials. The main contribution of this paper is to present a critical literature review on this important subject that was performed using all the related articles published in the past Frontiers in Education Conferences.

Within the tradition of Digital Game courses, the use and teaching of programming is seen as essential for students to work on skills and gain knowledge to develop an artifact. However, for students at a more basic level, this can be a major difficulty considering the complexity of understanding a language they have never seen before. Narrative, sound, and music are critical elements in game creation and can be used to teach programming concepts in a more intuitive and engaging way. Recently, the “Creative Experience: Text-Based Game Development” module has strengthened the use of narrative as a way of teaching programming concepts. Narrative creates contexts and stories for players, while sound provides feedback and guidance during the game. According to Kelleher (2006), the use of games to teach programming is an effective form of learning because it allows students to experiment with concepts in a fun and playful way and develops important skills such as problem-solving, critical thinking, and teamwork. The use of narrative, sound, and music in games can be an effective way to teach complex programming concepts clearly and engagingly. This is an article that addresses the strategies created in the digital game module “Creative Experience: Text-Based Game Development” exploring narrative, sound, and music as learning elements for programming and comparing the results within the period of two course offerings, one with only programming teaching and another with the introduction of narrative and sound as the basis for constructing this knowledge. It is hoped that these teaching strategies can contribute to the development of valuable skills in the field of education and programming.

Researchers dealing with real-world data - such as in the healthcare domain - tend to face class imbalance issues. More specifically, publicly available datasets containing Chest X-Ray (CXR) of Pneumonia diseases (including COVID-19) usually have an imbalanced class distribution. This dataset imbalance causes automatic diagnosis systems to classify majority classes with much more accuracy than the minority ones. Several resampling algorithms were proposed in the past to deal with the class imbalance issue. Hierarchical classifiers have also been proposed to increase the predictive performance of classifiers, but there is little research in the literature verifying if using existing resampling algorithms with hierarchical classifiers are a good alternative to improve classification performance. This work proposes an experimental classification schema to investigate the effectiveness of using resampling algorithms in the identification of COVID-19 and other types of Pneumonia through CXR images. The proposed schema uses resampling algorithms to rebalance the class distribution, in a Local Hierarchical Classification scenario. The experimental evaluation, which is supported by inferential statistical analysis, showed that using specific resampling algorithms with Local Hierarchical Classifiers brings a statistically significant increase to the macro-averaged Fl-Score, and improves the predictive performance for the minority classes.

We are currently experiencing a worldwide health problem known as the coronavirus pandemic, many researchers are looking to help in any way they can to deal with the pandemic and the problems caused by it. In the context of machine learning research, it is possible to develop methods to assist with the screening of patients using different types of exams and machine learning techniques. In this paper, we investigate the use of different features selection methods with different classifiers to the task of covid-19 (and other five pathologies and healthy lungs) identification in chest x-rays images. The analysis of the experimental results shows that the application of feature selection methods can improve the detection of coronavirus as well as other pathologies.

This Innovate Practice Full Paper presents how a group of teachers at our university have used Discord as an extension of the emergency remote teaching classroom due to the challenges imposed by the pandemic, caused by the coronavirus, for educators and students worldwide. Novel teaching and learning strategies, and several online tools, have been used to teach different subjects. In this paper we present how some teachers in a game development undergraduate course from the Polytechnic School have used the Discord communication tool as an extension of our virtual classroom. This paper presents some of the teaching strategies we have used with Discord as well as the students' perception and feedback about the use of this tool.

Contribution: This innovative practice full paper presents a novel approach based on active learning pedagogical strategies and entrepreneurship programmes to teach entrepreneurship to computer science and engineering students. The proposed approach has been tailored for undergraduate classes in order to provide the students with the initial stages of developing a technological startup. Background: Within the computer science and engineering literature, there are some interesting works that deal with entrepreneurial education in different levels and concepts. However, there has not been a paper describing a hands-on approach that makes them experience the important initial stages of a startup. Intended Outcomes: By the end of the course the students will be able to: (1) Validate whether or not their idea can become a business; (2) Build a MVP (Minimum Viable Product) for their solution; (3) Elaborate a Business Model Canvas (BMC); (4) Present their solution and BMC to potential investors. Application Design: There are some very interesting approaches to teach entrepreneurship in programmes outside of the university, such as the startup weekend, which has created several successful startups. Findings: Before taking this module the majority of the students did not have any contact with any type of entrepreneurship education programme or startup related event such as hackathons. While taking the module, several students attended related events. Some of them got involved with other university entrepreneurship initiates.

This Research Full Paper presents a computational hierarchical clustering approach to identify groups of college students from a Computer Science 1 (CS1) course that need extra help with the programming content. We first defined a set of features that characterize the student's programming skills in a CS1 course. Next, we applied a hierarchical clustering algorithm to bring together the students with similar skills by analyzing the source code they developed for different tasks. Finally, we evaluated the quality of the model and analyzed the different clusters. We processed a total of 630 source code tasks, for which our results indicate the formation of three clusters. We have found that one of the clusters has a large number of students with possible difficulties in programming. The other two clusters, although having different coding behaviors, reached high levels of knowledge about the contents taught. We conclude that features extracted from the students source codes can be used to group students into clusters that indicate their performance trends throughout the course.

In this study we classified the songs of the Latin Music Mood Database (LMMD) according to their emotion using two approaches: single-step classification, which consists of classifying the songs by emotion, valence, arousal and quadrant; and multistep classification, which consists of using the predictions of the best valence and arousal classifiers to classify quadrants and the best valence, arousal and quadrant predictions as features to classify emotions. Our hypothesis is that breaking the emotion classification in smaller problems would reduce complexity and improve results. Our best single-step emotion and valence classifiers used multimodal sets of features extracted from lyrics and audio. Our best arousal classifier used features extracted from lyrics and SMOTE to mitigate the dataset imbalance. The proposed multistep emotion classifier, which uses the predictions of a multistep quadrant classifier, improved the single-step classifier performance, reaching 0.605 of mean f-measure. These results show that using valence, arousal, and consequently, quadrant information can improve the prediction of specific emotions.

This Research Full Paper presents a review of recent studies on SIGCSE about teaching programming (CS1) and data structures (CS2) for university students in computer science courses. Our main contribution is the identification of three categories and their respective subcategories for teaching programming: (i) characterization of contents, (ii) identification of pedagogical strategies and (iii) grouping of support tools. Context: Technology is increasingly more present in modern life. This demands qualified professionals in algorithmic thinking and coding. However, teaching programming is a challenging task, with frequently high dropout and failure rates. Because of this there is a large amount of scientific literature about methods and tools for teaching this topic. A systematic review to compile existing techniques can be very useful for teacher and professionals interested in the area. Objective: The goal of this paper is to identify and discuss the recent approaches used in teaching algorithms and data structures presented in the last five years of SIGCSE. Thus, we investigated the following research question: What are the recent approaches to teaching programming and data structures? Method: We conducted a review of the primary researches published in SIGCSE Technical Symposium during 2014-2018. All primary studies published in this period were evaluated according to the inclusion criteria. The studies related to our objective were completely read, passed through the extraction of data and classified according to the research questions. Results: In the end of the selection phase, 60 papers were used to conduct this research. Our results were classified into three categories: (i) contents were identified in 60 papers. We found that CS1 is predominant, it is present in 76.7% of papers. Whereas the teaching of CS2 is present in only 38.3% of the papers; (ii) teaching strategies were identified in 32 papers. Active learning approaches are usually the target of these studies. It should be noted that pair programming was present in 18.3% of these papers; finally (iii) 28 papers were found about support tools and most of them are focused on visualization and animation. Conclusion: Effective pedagogical approaches for teaching CS1 and CS2 are of great concern for educators around the globe. There is also a need for more tools that support the teaching of these disciplines.

Video as well as music are potent means to convey emotions. However, despite their importance in several applications, few works deal with the issue of emotion classification in videos. The main reason is possibly the lack of available databases. In this work we extend the CAL500 database by including music videos, since the CAL500 was originally proposed as an audio-only database. The main rationale here is that the music videos must be official as they were developed to convey the same emotion as the song. After adapting the database, we have extracted audio and video features to perform our computational experiments. Our main result is that there is a complementarity between the audio and video features as the best result was achieved using their combination.

In this work we present a comprehensive set of experiments aiming to perform music genre classification using learned and handcrafted features plus the fusion of them. Handcrafted features were obtained from the audio signal itself, lyrics, chords and spectrogram images extracted from the audio. The rationale behind this investigation is based on the assumption that one can find some complementarity between classifiers created from these different resources. The experimental protocol was conducted on the Brazilian Music Dataset using the artist filter restriction and they confirm the power of non-handcrafted features to perform audio classification tasks. The experimental results have shown a significant complementarity among the handcrafted features for which the evaluated fusion strategies allowed an improvement in the classification accuracy up to 4 percent points. On the other hand, the fusion of learned and handcrafted features provided similar accuracy than the best individual CNN (0.7815).

This Innovative Practice Full Paper is grounded in the theme of teaching algorithms and programming to students with Special Educational Needs and Disabilities (SENDs) which is an increasingly difficult task for educators. The inclusion of SENDs in this context, makes the teaching task even more challenging, since in addition to the students' physical limitations, there is also the non-preparation of the teachers for specialized education. Among the SENDs, the students with visual impairment present one of the biggest obstacles in teaching algorithms. Considering the previously described context, this work proposes a method that uses Physical Flowcharts and Screen Readers in order to assist teaching algorithms to Visually Impaired students. To investigate the effectiveness of our method, a study was carried out in a Visually Impaired group of five students using the proposed method. The results of this study indicate that our method is promising and may be used in the future to help students with this kind of special needs.

In this Innovative Practice paper we describe our experience on using an updated methodology based on the paper “Music Education Meets Computer Science and Engineering” with a focus on pre-university female students. The main differences from this work are: (1) We taught the girls how to program and how to play the music instruments in parallel, having two classes per week, one for each subject instead of weekly meetings; (2) We started the project by using some of the games and puzzles from lightbot.com and code.org and then move to the Greenfoot Initial Learning Environment; (3) We asked the students to survey some music games related to their music instrument and design three potential games that they would like to develop using the music instruments; (4) They had to present and pitch the projects they designed to the undergraduate and graduate students of our computer music technology laboratory. Given the time constrains of the project, we had to help them to select the most feasible idea; (5) We applied a questionnaire to capture the perception of the students at the end of the project. It should be noted that all students successfully developed their music-based games using the digital music instruments and the Greenfoot Initial Learning Environment. In total four games were developed: (1) A game that allows drummers to perform ear training where they list to a particular drumming sequence and must replicate the sequence; (2) A game that allows guitar players to practice the execution of musical scales; (3) A game that allows keyboard players to practice their sight reading based on the simultaneous of two or more notes at the same time; (4) A game that allows bass players to practice their sight reading. It should be noted that by participating in the project one of the students acquired the skills and confidence to join the school and church bands.

The task of learning from imbalanced datasets has been widely investigated in the binary, multi-class and multilabel scenarios. Although this problem also affects hierarchical datasets, to the best of our knowledge, there are no works in the literature that deal with imbalanceness in hierarchical contexts. In this paper we propose metrics to measure "how imbalanced" is a Hierarchical Multi-Label Dataset, in addition to an approach to deal with this imbalanceness using Multi-Label resampling techniques. The proposed technique is based on the conversion of the dataset labels to a strictly multi-label format, applying wellknown multi-label resampling techniques and then converting the dataset back to its hierarchical taxonomy. The experimental evaluation over a highly imbalanced Music Genre Recognition dataset achieved promising results, with an increase of 0.2337 in the Avg-AUROC metric in relation to the original dataset.

In this work we address the automatic music genre classification as a pattern recognition task. The content of the music pieces were handled in the visual domain, using spectrograms created from the audio signal. This kind of image has been successfully used in this task since 2011 by extracting handcrafted features based on texture, since it is the main visual attribute found in spectrograms. In this work, the patterns were described by representation learning obtained with the use of convolutional neural network (CNN). CNN is a deep learning architecture and it has been widely used in the pattern recognition literature. Overfitting is a recurrent problem when a classification task is addressed by using CNN, it may occur due to the lack of training samples and/or due to the high dimensionality of the space. To increase the generalization capability we propose to explore data augmentation techniques. In this work, we have carefully selected strategies of data augmentation that are suitable for this kind of application, which are: adding noise, pitch shifting, loudness variation and time stretching. Experiments were conducted on the Latin Music Database (LMD), and the best obtained accuracy overcame the state of the art considering approaches based only in CNN.

Image identification of animals is mostly centred on identifying them based on their appearance, but there are other ways images can be used to identify animals, including by representing the sounds they make with images. In this study, the authors present a novel and effective approach for automated identification of birds and whales using some of the best texture descriptors in the computer vision literature. The visual features of sounds are built starting from the audio file and are taken from images constructed from different spectrograms and from harmonic and percussion images. These images are divided into sub‐windows from which sets of texture descriptors are extracted. The experiments reported in this study using a dataset of Bird vocalisations targeted for species recognition and a dataset of right whale calls targeted for whale detection (as well as three well‐known benchmarks for music genre classification) demonstrate that the fusion of different texture features enhances performance. The experiments also demonstrate that the fusion of different texture features with audio features is not only comparable with existing audio signal approaches but also statistically improves some of the stand‐alone audio features. The code for the experiments will be publicly available at https://www.dropbox.com/s/bguw035yrqz0pwp/ElencoCode.docx?dl=0 .

Music Genre Recognition (MGR) is one of the most important forms of music organization and has been an important research subject. Most MGR methods consider the audio content itself instead of the songs meta-data, such as its chords. However, the audio is not always available due to copyright issues, which makes the use of meta-data for MGR an important task. The main objective of this paper is to propose a new method for MGR using simplified chords sequences. We also propose a new public dataset with 8,994 songs containing audio and chords features from six different genres in order to evaluate our method. The experimental results reached an accuracy rate of 56.13% using only the chords sequence feature with the Random Forest classifier, and 78.40% combining audio and chords features with the SVM classifier.

With the steady growth of the Internet many search websites have emerged, however most of these websites only allow you to perform the search using a textual interface. A more intuitive method for the user would be to use melody snippets of a song to perform this search. In this paper, we present a two-stage query-by-playing (QBP) system using symbolic representations where from a melody snippet of a song it is performed the retrieval of a set of similar songs. On the first stage we perform the alignments and calculate its result rank, which may have draws. Stage two was created to help solving this problem, where a classifier-based filter is applied on the results, reducing the number of draws. A new methodology for evaluating QBP is presented, aiming to understand the impact of the number of notes in the query and some alternatives to verify the robustness of the methods considering wrong and missing notes. In addition, a new method based on machine learning (ML) is presented to filter the results. In the experiments this method always improves the performance of queries regardless of the noise.

Intrusion detection in computer networks is a important topic in information security. Due to numerous cases of security breaches that caused economic and social losses in recent years, this topic has been the subject of several studies in order to mitigate problems related to network intrusion and computer attacks. Information security systems have been using different techniques for network intrusion detection. However, with the development of communication mechanisms and consequently with the increase in data traffic, some techniques used for intrusion detection lost their information processing capability. The emergence of new forms of attacks on computer systems also contribute to the depreciation of some of the existing tools. In this scenario, new techniques capable of processing large amounts of information and that perform proactive discovery of new attack vectors are necessary. This paper presents a study on the use of a data stream mining technique known as Hoeffding Adaptive Tree to create a predictive model for network intrusion detection. The experiments performed in this work show the effectiveness of this technique when applied to a database of computer network attacks.

Previous artificial intelligence education research (DeNero and Klein, 2010) has used the classic video game Pacman to teach introductory artificial intelligence concepts. One of the advantages of the work proposed in (DeNero and Klein, 2010) is that the same framework, in this case using the video game Pacman, can be used for different student assignments covering different artificial intelligence algorithms and methods. The issue of how to use the same practical framework is an important one, because if students have to learn a different framework for every assignment they will often feel discouraged. For this reason, the main contribution of this paper is to present a practical assignment to teach students about genetic algorithm-based parameter optimization using the Pacman framework.

This paper presents a novel interdisciplinary approach to aid with the growing concern about how to showcase computer science and engineering degrees to pre-university level students. This novel approach is based on empowering students to create their own music-related game using real music instruments. In order to allow the students to program their games we have used the Greenfoot Introductory Learning Environment to teach object-oriented programming skills with the java programming language. We have also taught the students music theory and practice. All these skills were required to allow the students to successfully develop their musical games that interact with the music instruments.

Introductory artificial intelligence undergraduate classes often introduce different search methods using different search algorithms. In this context one of algorithms that is often taught, is the minimax algorithm which is used in adversarial games where you want to minimize your opponent's chance of winning while maximizing your chance of winning. Different instructors use different games to make the students implement the minimax algorithm such as Checkers, Othello or Chess. However, one common problem with this assignment is that the students often spend more time implementing the game itself rather than the artificial intelligence techniques in the game. For this reason, in this paper we present a java-based open source Othello framework that was designed to be used in artificial intelligence undergraduate classes. Our framework has several features that help the students to focus on the development of the artificial intelligence aspects of the game, rather than developing the game itself. One particular feature of the framework is that it has a method that returns the list of valid moves given the current state of the game board and which player is going to make the next move. With this method, the students can focus on how to evaluate the different states using several heuristic functions and implementing the minimax algorithm. Another feature of the framework is the graphical user interface and the HumanPlayer class that allows the students to play against their own code. This feature is important as it allows the students to not only debug their codes but also to evaluate the effectiveness of their implemented heuristics. Another aspect of the framework is that it allows to set up a tournament of the codes developed by the students. The tournament can be organized in two modes. In the first mode every AI developed by one student plays against the AI developed by every other student. In the second mode, each student developed code is paired against another student developed code and only the winner plays against the winner of another pairing until there is only one winner left. An analysis of the framework in our artificial intelligence undergraduate computer engineering classes shows that it properly supports the student learning and the tournament mode also challenges them to create the best AI for Othello as they can.

In this paper a novel approach for automatic bird species classification is described. The proposed strategy is based on features taken from the textural content of spectrogram images of bird vocalizations. We show how several texture descriptors can be used for representing the spectrograms. The following approaches are tested here with spectrograms for the first time: Local Ternary Phase Quantization, Heterogeneous Auto-Similarities of Characteristics, and an ensemble of variants of Local Binary Pattern Histogram Fourier. Combining this set of descriptors greatly increases classification performance and markedly improves previous ensembles of texture descriptors used for describing a spectrogram. Moreover, a further improvement is obtained when the texture descriptors are combined with the acoustic features. SVM classifiers are used in the classification step, with final results computed using 10-fold cross-validation. For a fair comparison with other methods in the literature, the experiments are performed on a benchmark database composed of 46 bird species used for this classification task. The best accuracy rate obtained is about 94.5%. The MATLAB code we used is publicly available to other researchers for future comparisons, as well as the database used in the experiments.

The problem of converting a music in standard music notation (music sheet) to the alternative notation of guitar tablature is known as transcription. The process of transcription consists of indicating where each note from the original music sheet needs to be played in the guitar, i.e. which string and fret of the guitar that needs to be played to produce a particular note. However, considering that each note can be played in different positions of the guitar fretboard, this is not a straightforward process, and can be classified as a combinatorial optimization problem. For this reason, we have employed a comparative study of different algorithms: A-star, genetic algorithms (GA), genetic algorithms based on subpopulations (GA-SP), ant colony optimization (ACO) and differential evolution (DE). It was also included heuristics based on local search 2-opt and 3-opt in the approaches GA, GA-SP and DE. The experimental results with a dataset of 87 musics indicated that the approaches ACO, GA-SP with 2-opt and GA with 2-opt reached the best performance. Also, the results obtained with each approach were statistically compared using ANOVA test with post hoc Tukey.

This article describes the automatic classification of emotions through five novel feature descriptors based on the "cifras" of the songs. A "cifra" is a document that contains the harmonic structure and harmonic progression a song. In order to evaluate the proposed feature sets, we have created a novel database, namely the Latin Music Cifras Mood Database (LMCMD). This database was created manually by the authors by querying different websites for the "cifras" of songs.

The identification of bird species from their audio recorded songs are nowadays used in several important applications, such as to monitor the quality of the environment and to prevent bird-plane collisions near airports. The complete identification cycle involves the use of: (a) recording devices to acquire the songs, (b) audio processing techniques to remove the noise and to select the most representative elements of the signal, (c) feature extraction procedures to obtain relevant characteristics, and (d) decision procedures to make the identification. The decision procedures can be obtained by Machine Learning (ML) algorithms, considering the problem in a standard classification scenario. One key element is this cycle is the selection of the most relevant segments of the audio for identification purposes. In this paper we show that the use of short audio segments with high amplitude - called pulses in our work - outperforms the use of the complete audio records in the species identification task. We also show how these pulses can be automatically obtained, based on measurements performed directly on the audio signal. The employed classifiers are trained using a previously labeled database of bird songs. We use a database that contains bird song recordings from 75 species which appear in the Southern Atlantic Coast of South America. Obtained results show that the use of automatically obtained pulses and a SVM classifier produce the best results, all the necessary procedures can be installed in a dedicated hardware, allowing the construction of a specific bird identification device.

The automatic music genre classification task is an active area of research in the field of Music Information Retrieval. In this paper we use two different symbolic feature sets for genre classification and combine them using an early fusion approach. Our results show that early fusion achieves better classification accuracy than using any of the individual feature sets. Furthermore, when compared with some of the state of the art approaches using the same experimental conditions, early fusion of symbolic features is ranked the second best method.

In this paper we address the task of hierarchical bird species identification from audio recordings. We evaluate three types of approaches to deal with hierarchical classification problems: the flat classification approach, the local-model per parent node classifier approach and the global-model hierarchical-classification approach. For the flat and local-model classification approach we employ the classic Naive Bayes algorithm. For the global-model approach we use the Global Model Naive Bayes (GMNB) algorithm. As in the classical Naive Bayes, the algorithm computes prior probabilities and likelihoods, but these computations take into account the hierarchical classification scenario: it assumes that any example which belongs to a given class will also belong to all its ancestor classes. In the current application, the employed class hierarchy is the standard scientific taxonomy of birds used in Biology. In order to deal with the bird songs we obtain features by computing several acoustic quantities from intervals of the audio signal. We conduct three experiments in order to compare the three different approaches to the hierarchical bird species identification problem. Our experimental results show that the use of the GMNB hierarchical classification algorithm outperforms both the flat and local-model approaches (Using the Hierarchical F-measure metric), hence the use of a global-model approach (such as the GMNB) can be a feasible way to improve the classification performance for problems with a large number of classes.

Nowadays, the representation of many real word problems needs to use some type of relational model. As a consequence, information used by a wide range of systems has been stored in multi relational tables. However, from a data mining point of view, it has been a problem, since most of the traditional data mining algorithms have not been originally proposed to handle this type of data without discarding relationship information. Aiming to ameliorate this problem, we propose a hierarchical approach for handling relational data. In this approach the relational data is converted into a hierarchical structure (the main table as the root and the relations as the nodes). This hierarchical way to represent relational data can be used either for classification or clustering purposes. In this paper, we will use it in clustering algorithms. In order to do so, we propose a hierarchical distance metric to compute the similarity between the tables. In the empirical analysis, we will apply the proposed approach in two well-known clustering algorithms (k-means and agglomerative hierarchical). Finally, this paper also compares the effectiveness of our approach with one existing relational approach.

In this paper we focus on the automatic identification of bird species from their audio recorded song. Bird monitoring is important to perform several tasks, such as to evaluate the quality of their living environment or to monitor dangerous situations to planes caused by birds near airports. We deal with the bird species identification problem using signal processing and machine learning techniques. First, features are extracted from the bird recorded songs using specific audio treatment, next the problem is performed according to a classical machine learning scenario, where a labeled database of previously known bird songs are employed to create a decision procedure that is used to predict the species of a new bird song. Experiments are conducted in a dataset of recorded songs of bird species which appear in a specific region. The experimental results compare the performance obtained in different situations, encompassing the complete audio signals, as recorded in the field, and short audio segments (pulses) obtained from the signals by a split procedure. The influence of the number of classes (bird species) in the identification accuracy is also evaluated.

This paper deals with the automated bird species identification problem, in which it is necessary to identify the species of a bird from its audio recorded song. This is a clever way to monitor biodiversity in ecosystems, since it is an indirect non-invasive way of evaluation. Different features sets which summarize in different aspects the audio properties of the audio signal are evaluated in this paper together with machine learning algorithms, such as probabilistic, instance-based, decision trees, neural networks and support vector machines. Experiments are conducted in a dataset of recorded songs of three bird species. The experimental results compare the performance of the features sets and different classifiers showing that it is possible to obtain very promising results in the automated bird species identification problem.

Automatically inferring the function of unknown proteins is a challenging task in proteomics. There are two major problems in the task of computational protein function prediction, which are the choice of the protein representation and the choice of the classification algorithm. There are several w ays of extracting features from a protein, and the choice of the feature representation might be as important as the choice of the classification algorithm. These problems are aggravated in the case of hierarchical protein function prediction, where a hierarchy of classifiers is built and each of those classifiers' construction has to consider the aforementioned selection problems. In this paper we address these problem by employing three alternative selective hierarchical classification approaches: (a) selecting the best classifier given a fixed representation; (b) selecting the best representation given a fixed classifier; and (c) selecting the best classifier and representation simultaneously, in a synergistic fashion. The analysis of the results have shown that the selective representation approach is almost always ranked number 1 when compared against the different fixed representations and that the use of the selective classifier approach is not able to surpass using only the best classifier for the target problem.

In this paper we propose a new global-model approach for hierarchical classification, where a single global classification model is built by considering all the classes in the hierarchy - rather than building a number of local classification models as it is more usual in hierarchical classification. The method is an extension of the flat classification algorithm naive Bayes. We present the extension made to the original algorithm as well as its evaluation on eight protein function hierarchical classification datasets. The achieved results are positive and show that the proposed global model is better than using a local model approach.

This paper presents two novel hierarchical classification methods which are extensions of a previously proposed selective classifier top-down approach, which consists of selecting - during the training phase - the best classifier at each node of a classifier tree. More precisely, we propose two novel selective top-down hierarchical methods. First, a method that selects the best feature set instead of the best classifier. Secondly, a method that selects both the best classifier and the best representation simultaneously. These methods are evaluated on the task of hierarchical music genre classification using four different types of feature sets extracted from each song and four classifiers.

This paper presents the results of the application of a feature selection procedure to an automatic music genre classification system. The classification system is based on the use of multiple feature vectors and an ensemble approach, according to time and space decomposition strategies. Feature vectors are extracted from music segments from the beginning, middle and end of the original music signal (time decomposition). Despite being music genre classification a multi-class problem, we accomplish the task using a combination of binary classifiers, whose results are merged in order to produce the final music genre label (space decomposition). As individual classifiers several machine learning algorithms were employed: naive-Bayes, decision trees, support vector machines and multi-layer perceptron neural nets. Experiments were carried out on a novel dataset called Latin music database, which contains 3,227 music pieces categorized in 10 musical genres. The experimental results show that the employed features have different importance according to the part of the music signal from where the feature vectors were extracted. Furthermore, the ensemble approach provides better results than the individual segments in most cases.

This paper presents a novel approach to the task of automatic music genre classification which is based on multiple feature vectors and ensemble of classifiers. Multiple feature vectors are extracted from a single music piece. First, three 30-second music segments, one from the beginning, one from the middle and one from end part of a music piece are selected and feature vectors are extracted from each segment. Individual classifiers are trained to account for each feature vector extracted from each music segment. At the classification, the outputs provided by each individual classifier are combined through simple combination rules such as majority vote, max, sum and product rules, with the aim of improving music genre classification accuracy. Experiments carried out on a large dataset containing more than 3,000 music samples from ten different Latin music genres have shown that for the task of automatic music genre classification, the features extracted from the middle part of the music provide better results than using the segments from the beginning or end part of the music. Furthermore, the proposed ensemble approach, which combines the multiple feature vectors, provides better accuracy than using single classifiers and any individual music segment.