Pablo M. Granitto

Neural network ensembles: evaluation of aggregation algorithms

Ensembles of artificial neural networks show improved generalization capabilities that outperform those of single networks. However, for aggregation to be effective, the individual networks must be as accurate and diverse as possible. An important problem is, then, how to tune the aggregate members in order to have an optimal compromise between these two conflicting conditions. We present here an extensive evaluation of several algorithms for ensemble construction, including new proposals and comparing them with standard methods in the literature. We also discuss a potential problem with sequential aggregation algorithms: the non-frequent but damaging selection through their heuristics of particularly bad ensemble members. We introduce modified algorithms that cope with this problem by allowing individual weighting of aggregate members. Our algorithms and their weighted modifications are favorably tested against other methods in the literature, producing a sensible improvement in performance on most of the standard statistical databases used as benchmarks.

Weed seeds identification by machine vision

Abstract The implementation of new methods for reliable and fast identification and classification of seeds is of major technical and economical importance in the agricultural industry. As in ocular inspection, the automatic classification of seeds should be based on knowledge of seed size, shape, color and texture. In this work, we assess the discriminating power of these characteristics for the unique identification of seeds of 57 weed species. Using the performance of a naive Bayes classifier as selection criterion, we identified a nearly optimal set of 12 (six morphological+four color+two textural) seed characteristics to be used as classification parameters. We found that, as expected, size and shape characteristics have larger discriminating power than color and textural ones. However, all these features are required to reach an identification performance acceptable for practical applications. In spite of its simplicity, the naive Bayes classifier reveals itself surprisingly good for the identification of seed species. This might be due to the careful selection of the feature set, leading to nearly independent parameters as assumed by this method. We also found that, using the same feature set, a more sophisticated classifier based on an artificial neural network committee performs only slightly better than this simple Bayesian approach.

Rapid characterization of dry cured ham produced following different PDOs by proton transfer reaction time of flight mass spectrometry (PTR-ToF-MS)

In the present study, the recently developed proton transfer reaction time of flight mass spectrometry (PTR-ToF-MS) technique was used for the rapid characterization of dry cured hams produced according to 4 of the most important Protected Designations of Origin (PDOs): an Iberian one (Dehesa de Extremadura) and three Italian ones (Prosciutto di San Daniele, Prosciutto di Parma and Prosciutto Toscano). In total, the headspace composition and respective concentration for nine Spanish and 37 Italian dry cured ham samples were analyzed by direct injection without any pre-treatment or pre-concentration. Firstly, we show that the rapid PTR-ToF-MS fingerprinting in conjunction with chemometrics (Principal Components Analysis) indicates a good separation of the dry cured ham samples according to their production process and that it is possible to set up, using data mining methods, classification models with a high success rate in cross validation. Secondly, we exploited the higher mass resolution of the new PTR-ToF-MS, as compared with standard quadrupole based versions, for the identification of the exact sum formula of the mass spectrometric peaks providing analytical information on the observed differences. The work indicates that PTR-ToF-MS can be used as a rapid method for the identification of differences among dry cured hams produced following the indications of different PDOs and that it provides information on some of the major volatile compounds and their link with the implemented manufacturing practices such as rearing system, salting and curing process, manufacturing practices that seem to strongly affect the final volatile organic profile and thus the perceived quality of dry cured ham.

Automatic classification of legumes using leaf vein image features

In this paper, a procedure for segmenting and classifying scanned legume leaves based only on the analysis of their veins is proposed (leaf shape, size, texture and color are discarded). Three legume species are studied, namely soybean, red and white beans. The leaf images are acquired using a standard scanner. The segmentation is performed using the unconstrained hit-or-miss transform and adaptive thresholding. Several morphological features are computed on the segmented venation, and classified using four alternative classifiers, namely support vector machines (linear and Gaussian kernels), penalized discriminant analysis and random forests. The performance is compared to the one obtained with cleared leaves images, which require a more expensive, time consuming and delicate procedure of acquisition. The results are encouraging, showing that the proposed approach is an effective and more economic alternative solution which outperforms the manual experts recognition. HighlightsWe develop an automatic procedure to classify legume species using scanned leaves.The method is based exclusively on the analysis of the leaf venation images.We analyze the advantages over the usage of cleared leaves.Different state-of-the-art classifiers are compared.The proposed method outperforms human expert classification.

PTR-ToF-MS and data mining methods: a new tool for fruit metabolomics

Proton Transfer Reaction-Mass Spectrometry (PTR-MS) in its recently developed implementation based on a time-of-flight mass spectrometer (PTR-ToF-MS) has been evaluated as a possible tool for rapid non-destructive investigation of the volatile compounds present in the metabolome of apple cultivars and clones. Clone characterization is a cutting-edge problem in technical management and royalty application, not only for apple, aiming at unveiling real properties which differentiate the mutated individuals. We show that PTR-ToF-MS coupled with multivariate and data mining methods may successfully be employed to obtain accurate varietal and clonal apple fingerprint. In particular, we studied the VOC emission profile of five different clones belonging to three well known apple cultivars, such as ‘Fuji’, ‘Golden Delicious’ and ‘Gala’. In all three cases it was possible to set classification models which can distinguish all cultivars and some of the clones considered in this study. Furthermore, in the case of ‘Gala’ we also identified estragole and hexyl 2-methyl butanoate contributing to such clone characterization. Beside its applied relevance, no data on the volatile profiling of apple clones are available so far, our study indicates the general viability of a metabolomic approach for volatile compounds in fruit based on rapid PTR-ToF-MS fingerprinting.

Deep learning for plant identification using vein morphological patterns

Display Omitted Deep convolutional neural network (CNN) for plant identification focusing on leaf vein patterns.No task-specific feature extractors needed.Improved the state of the art accuracy on a legume species recognition task.Visualization of relevant vein patterns. We propose using a deep convolutional neural network (CNN) for the problem of plant identification from leaf vein patterns. In particular, we consider classifying three different legume species: white bean, red bean and soybean. The introduction of a CNN avoids the use of handcrafted feature extractors as it is standard in state of the art pipeline. Furthermore, this deep learning approach significantly improves the accuracy of the referred pipeline. We also show that the reported accuracy is reached by increasing the model depth. Finally, by analyzing the resulting models with a simple visualization technique, we are able to unveil relevant vein patterns.

Effect of the pig rearing system on the final volatile profile of Iberian dry-cured ham as detected by PTR-ToF-MS

The volatile compound profile of dry-cured Iberian ham lean and subcutaneous fat from pigs fattened outdoors on acorn and pasture (Montanera) or on high-oleic concentrated feed (Campo) was investigated by proton transfer reaction time-of-flight mass spectrometry. In addition to the usual proton transfer ionization the novel switchable reagent ions system was implemented which allows the use of different precursor ions (H(3)O(+), NO(+) and O(2)(+)). The analysis of the lean and subcutaneous fat volatile compounds allowed a good sample discrimination according to the diet. Differences were evident for several classes of compounds: in particular, Montanera hams showed higher concentrations of aldehydes and ketones and lower concentrations of sulfur-containing compounds compared to Campo hams. The use of NO(+) as precursor ion confirmed the results obtained with H(3)O(+) in terms of classification capability and provides additional analytical insights.

Aggregation algorithms for neural network ensemble construction

How to generate and aggregate base learners to have optimal ensemble generalization capabilities is an important questions in building composite regression/classification machines. We present here an evaluation of several algorithms for artificial neural networks aggregation in the regression settings, including new proposals and comparing them with standard methods in the literature. We also discuss a potential problem with sequential algorithms: the non frequent but damaging selection through their heuristics of particularly bad ensemble members. We show that one can cope with this problem by allowing individual weighting of aggregate members. Our algorithms and their weighted modifications are favorably tested against other methods in the literature, producing a performance improvement on the standard statistical databases used as benchmarks.

Selecting diverse members of neural network ensembles

Ensembles of artificial neural networks have been used as classification/regression machines, showing improved generalization capabilities that outperform those of single networks. However, it has been recognized that for aggregation to be effective the individual network must be as accurate and diverse as possible. An important problem is, then, how to choose the aggregate members in order to have an optimal compromise between these two conflicting conditions. We propose here a new method for selecting members of regression/classification ensembles that leads to small aggregates with few but very diverse individual predictors. Using artificial neural networks as individual learners, the algorithm is favorably tested against other methods in the literature, producing a remarkable performance improvement on the standard statistical databases used as benchmarks. In addition, and as a concrete application, we study the sunspot time series and predict the remaining of the current cycle 23 of solar activity.

Predictions of the maximum amplitude for solar cycle 23 and its subsequent behavior using nonlinear methods

Two nonlinear methods are employed for the prediction of the maximum amplitude for solar cycle 23 and its declining behavior. First, a new heuristic method based on the second derivative of the (conveniently smoothed) sunspot data is proposed. The curvature of the smoothed sunspot data at cycle minimum appears to correlate (R ≃ 0.92) with the cycles later-occurring maximum amplitude. Secondly, in order to predict the near-maximum and declining activity of solar cycle 23, a neural network analysis of the annual mean sunspot time series is also performed. The results of the present study are then compared with some other recent predictions.