Camila Maione’s research while affiliated with Universidade Federal de Goiás and other places

Soil and dirt fragments are easily transferred from the ground to objects such as clothing, shoes, skin, nails, and tires. The elemental analysis of these samples involved in crimes can be an important source of information for forensic scientists because they can present substantial evidence by creating links between victims, suspects, crime scenes and other relevant actors or places. In this work we present a promising new approach for the study of soil samples, using data mining techniques applied to the elemental fingerprints of soil. We experimented on soil samples obtained from southeast Oregon and northern Nevada, two neighboring states in the United States that have similar geological characteristics while also displaying some specific differences. The chemical composition of soil and sediments samples were determined by the use of inductively coupled plasma-mass spectrometry (ICP-MS). Thirty-three elements were analyzed, and we used their concentrations to conduct the analysis. Cluster analysis was performed employing the K-means clustering algorithm. We found three clusters that showed interesting chemical patterns. In order to investigate the most significant chemical elements that distinguish the clusters, we employed the Correlation-Based Feature Selection (CFS) algorithm. Lastly, we developed a classification model based on support vector machine (SVM), which can predict in which of the found clusters an arbitrary soil sample would fall with a 99% prediction accuracy when all 32 variables were used for training the model, and a 95% prediction accuracy when only the 10 most relevant elements were used for training the model. Following this methodology, forensic scientists and experts would be able to establish profiles of soil samples extracted from the crime scene and nearby regions, and use classification models to predict which of these profiles an arbitrary soil sample found on the subjects involved in the crime would be associated with.

Lettuce (Lactuca sativa) is the main leafy vegetable produced in Brazil. Since its production is widespread all over the country, lettuce traceability and quality assurance is hampered. In this study, we propose a new method to identify the geographical origin of Brazilian lettuce. The method uses a powerful data mining technique called support vector machines (SVM) applied to elemental composition and soil properties of samples analyzed. We investigated lettuce produced in São Paulo and Pernambuco, two states in the southeastern and northeastern regions in Brazil, respectively. We investigated efficiency of the SVM model by comparing its results with those achieved by traditional linear discriminant analysis (LDA). The SVM models outperformed the LDA models in the two scenarios investigated, achieving an average of 98 % prediction accuracy to discriminate lettuce from both states. A feature evaluation formula, called F–score, was used to measure the discriminative power of the variables analyzed. The soil exchangeable cation capacity, soil contents of low crystalized Al and Zn content in lettuce samples were the most relevant components for differentiation. Our results reinforce the potential of data mining and machine learning techniques to support traceability strategies and authentication of leafy vegetables.

Because honey is a natural product synthetized by bees from secretions of various flowers and plants, its composition and properties are determined by those unique floral origins. Aside from honey’s sweet and distinctive flavor, it can provide various human health benefits, which makes its market value favorable to those of other sweeteners. Consequently, honey has been a very common and easy target of adulteration for economic profit, making authenticity of honey a longtime concern for researchers and producers around the world. The oldest and most popular of these methods is melissopalynology, the analysis of pollen contained in honey, which can determine its floral origin. However, because this method is time consuming, requires a specialist, and is unable to detect fraudulent pollen contamination, several other methods for the determination of the floral and geographical origin of honey have been proposed in the scientific literature. Multivariate data analysis and machine learning are emerging areas that offer performance and cost advantages for extracting valuable information from raw data sets. They are capable of performing both exploratory and predictive analyses, which can help to uncover trends and hidden patterns within data. Most of the studies reviewed base their methods on atomic spectra and physicochemical properties as descriptive variables for honey. Sensorial data obtained from electronic tongue and nose, and color histograms of honey images also show high discriminative power for ascertaining honey origin. Principal component analysis (PCA), discriminant analysis (DA), and cluster analysis are the preferred techniques for performing exploratory and predictive analyses for the purpose of identifying origin. PCA and DA continue to be preferred due to their ease of application and interpretation, while machine learning algorithms are more complex to model. Nevertheless, both machine learning algorithms and PCA-DA models achieved excellent results in discriminating honey origin. Finally, a commonly observed tendency is the use of hybrid methods that combine multivariate data analysis and machine learning techniques, as each technique has its own strengths and weaknesses regarding uncovering and classifying patterns in honey data.

Ecstasy is an amphetamine-type substance that belongs to a popular group of illicit drugs known as “club drugs” whose consumption is rising in Brazil. The effects caused by this substance in the human organism are mainly psychological, including hallucinations, euphoria and other stimulant effects. The distribution of this drug is illegal, and effective strategies are required in order to detain its growth. One possible way to obtain useful information on ecstasy trafficking routes, sources of supply, clandestine laboratories and synthetic protocols is by its chemical components. In this paper, we present a data mining and predictive analysis for ecstasy tablets seized in two cities of São Paulo state (Brazil), Campinas and Ribeirão Preto, based on their chemical profile. We use the concentrations of 25 elements determined in the ecstasy samples by ICP-MS as our descriptive variables. We develop classification models based on support vector machines capable of predicting in which of the two cities an arbitrary ecstasy sample was most likely to have been seized. Our best model achieved a 81.59% prediction accuracy. The F-score measure shows that Se, Mo and Mg are the most significant elements that differentiate the samples from the two cities, and they alone are capable of yielding an SVM model which achieved the highest prediction accuracy.

This paper presents a data mining study and cluster analysis of social data obtained on small producers and family farmers from six country cities in Ceará state, northeast Brazil. The analyzed data involve demographic, economic, agriculture and food insecurity information. The goal of the study is to establish profiles for the small producer families that reside in the region and to identify relevant features which differentiate these profiles. Moreover, we provide an efficient data mining methodology for analysis of social data sets which is capable of handling its natural challenges, such as mixed variables and abundance of null values. We use the Silhouette method for the estimation of the best number of natural groups within the data, along with the Partitioning Around Medoids clustering algorithm in order to compute the profiles. The Correlation-Based Feature Selection method is used to identify which social criteria are the most important to differentiate the families from each profile. Classification models based on support vector machines, multilayer perceptron and decision trees were developed aiming to predict in which of the identified clusters an arbitrary family would be best fit. We obtained a good separation of the families into two clusters, and a multilayer perceptron model with approximately 93.5% prediction accuracy.

Rice is one of the most important staple foods around the world. Authentication of rice is one of the most addressed concerns in the present literature, which includes recognition of its geographical origin and variety, certification of organic rice and many other issues. Good results have been achieved by multivariate data analysis and data mining techniques when combined with specific parameters for ascertaining authenticity and many other useful characteristics of rice, such as quality, yield and others. This paper brings a review of the recent research projects on discrimination and authentication of rice using multivariate data analysis and data mining techniques. We found that data obtained from image processing, molecular and atomic spectroscopy, elemental fingerprinting, genetic markers, molecular content and others are promising sources of information regarding geographical origin, variety and other aspects of rice, being widely used combined with multivariate data analysis techniques. Principal component analysis and linear discriminant analysis are the preferred methods, but several other data classification techniques such as support vector machines, artificial neural networks and others are also frequently present in some studies and show high performance for discrimination of rice.

Chocolate is an appreciated food derived from cacao fruit. The flavonoids and minerals present in the chocolate have benefits to health, and some specific minerals are known to be toxic. Because of differences in their production systems, organic chocolate has a distinguishable pattern in mineral concentrations than conventional chocolate. Aiming for authenticity and study of the toxic elements in organic chocolate, we present in this work a simple method to classify organic chocolate samples based on elemental fingerprint profiling and multivariate data analysis. Thirty‐eight elements (toxic and essential) were determined in 36 chocolate samples (organic and conventional) by using inductively coupled plasma mass spectrometry to establish reference ranges and to identify differences in patterns of elements in both type of samples. Our results showed that Al, Zn, Mn, Cu, and Ba are the most present components for both types of chocolate, and higher concentrations of essential elements Fe, Zn, and Mg are found in conventional type, opposing the idea that organic food is rich in essential elements. Principal component analysis and linear discriminant analysis were used for multivariate data analysis, and sample differentiation was possible with 83% accuracy. We present a method to differentiate organic from conventional chocolate based on elemental fingerprint profiling and multivariate data analysis. Thirty‐eight elements (toxic and essential) were determined in samples by inductively coupled plasma—mass spectrometry. We found that Al, Zn, Mn, Cu, and Ba are the most commonly present elements for both types of chocolate, whereas higher concentrations of essential elements Fe, Zn, and Mg are found in the conventional samples.

Brazil is the world's largest producer of oranges. The Brazilian conventional citrus crop requires repeated application of agrochemicals to achieve satisfactory levels of productivity. The organic citriculture is an alternative production system, which is environmentally friendly and offers a safe food to consumers. However, it is difficult to determine if a food or plant was cultivated in organic or conventional system by just common observation, which makes the customers of organic food market vulnerable against fraudulent entrepreneurs. In this study we present a data mining approach for the study of Brazilian organic citrus leaves which can aid in the certification of authenticity of the citrus leaves. The elemental composition is determined by inductively coupled plasma – mass spectrometry (ICP-MS). We developed classification models based on support vector machines and artificial neural networks capable of predicting whether a citrus leaf is organic or conventional through analysis of the concentration levels of the fourteen chemical elements (Al, Ba, Co, Cr, Cs, Cu, Fe, Mg, Mn, Ni, Rb, Si, Sr and V) found in both types of leaves. Feature selection filter methods are used to determine the most relevant elements for the classification process. Our best model obtained was a support vector machine with approximately 88% prediction accuracy. The elements Mn, Mg, and Rb were evaluated as the most significant for the classification decision. This is the first paper which addresses the problem of classification of organic orange leaves based on chemical composition. The presented methodology is useful for attesting authenticity of organic citrus leaves and can be adapted for other organic food or substances.

The variations found in the elemental composition in ecstasy samples result in spectral profiles with useful information for data analysis, and cluster analysis of these profiles can help uncover different categories of the drug. We provide a cluster analysis of ecstasy tablets based on their elemental composition. Twenty-five elements were determined by ICP-MS in tablets apprehended by Sao Paulo's State Police, Brazil. We employ the K-means clustering algorithm along with C4.5 decision tree to help us interpret the clustering results. We found a better number of two clusters within the data, which can refer to the approximated number of sources of the drug which supply the cities of seizures. The C4.5 model was capable of differentiating the ecstasy samples from the two clusters with high prediction accuracy using the leave-one-out cross-validation. The model used only Nd, Ni, and Pb concentration values in the classification of the samples.