Agnieszka Kubik-Komar

Musical Instruments in Random Forest

This paper describes automatic classification of predominant musical instrument in sound mixes, using random forests as classifiers. The description of sound parameterization applied and methodology of random forest classification are given in the paper. Additionally, the significance of sound parameters used as conditional attributes is investigated. The results show that almost all sound attributes are informative, and random forest technique yields much higher classification results than support vector machines, used in previous research on these data.

Analysis of Recognition of a Musical Instrument in Sound Mixes Using Support Vector Machines

Experiments with recognition of the dominating musical instrument in sound mixes are interesting from the point of view of music information retrieval, but this task can be very difficult if the mixed sounds are of the same pitch. In this paper, we analyse experiments on recognition of the dominating instrument in mixes of same-pitch sounds of definite pitch. Sound from one octave (no. 4 in MIDI notation) have been chosen, and instruments of various types, including percussive instruments were investigated. Support vector machines were used in our experiments, and statistical analysis of the results was also carefully performed. After discussing the outcomes of these experiments and analyses, we conclude our paper with suggestions regarding directions of possible future research on this subject.

The effect of biotic and physical factors on the competitive ability of Rhizobium leguminosarum

BackgroundRhizobium leguminosarum bv. viciae (Rlv) is a soil bacterium which can form nitrogen-fixing symbiotic relationships with leguminous plants. Numerous rhizobial strains found in soils compete with each other. Competition can occur both during the saprophytic growth phase in the rhizosphere and inside plant tissues, during the symbiotic phase. Competition is important as it may affect the composition of rhizobial populations present in the soil and in the root nodules of plants.MethodologyWe examined the link between physiological traits and bacterial competitive ability in eighteen Rhizobium leguminosarum bv. viciae (Rlv) isolates during root nodule colonization using laboratory and field experiments. The competitive ability of R/v strains was measured as the percentage of root nodules colonized by gusA-tagged rhizobia in two types of host plants, peas and vetch.ResultsThe competitiveness of Rlv strains was significantly affected by soil type and the identity of the host plant. Of the eighteen bacterial traits examined in this study, the metabolic potential (number of utilized carbon and energy sources) and the responsiveness of nod genes to flavonoid activation were most important in affecting the competitive ability of Rlv strains. The amount of acylated homoserine lactones (AHL) produced by the strains was less important in influencing competitiveness. Finally, the preactivation of strains with flavonoids or the addition of AHL to gus-tagged Rlv strains did not significantly enhance competitiveness: of the gus-tagged inoculants in comparison to indigenous soil populations of vetch microsymbionts.ConclusionsThe competitiveness of Rlv strains is dependent upon numerous physiological traits. However, environmental factors such as soil type and the type of host plant may be even more important in affecting rhizobial competitiveness.

Application of analysis of variance and post hoc comparisons to studying the discriminative power of sound parameters in distinguishing between musical instruments

In this paper, the influence of the selected sound features on distinguishing between musical instruments is presented. The features were chosen basing on our previous research. Coherent groups of features were created on the basis of significant features, according to the parameterization method applied, in order to constitute small, homogenous groups. In this research, we investigate (for each feature group separately) if there exist significant differences between means of these features for the studied instruments. We apply analysis of variance along with post hoc comparisons in the form of homogeneous groups, defined by mean values of the investigated features for our instruments. If a statistically significant difference is found, then the homogenous group is established. Such a group may consist of only one instrument (distinguished by this feature), or more (instruments similar with respect to this feature). The results show which instruments can be best discerned by which features.

Spirulina enhances the viability of Lactobacillus rhamnosus E/N after freeze‐drying in a protective medium of sucrose and lactulose

Aims:  Response surface methodology (RSM) was used to optimize a protective medium for enhancing the viability of Lactobacillus rhamnosus E/N cells during lyophilization.

The plackett-burman design in optimization of media components for biomass production of Lactobacillus rhamnosus OXY.

The central composite design was developed to search for an optimal medium for the growth of Lactobacillus rhamnosus OXY. The effect of various media components, such as carbon sources, simple and complex nitrogen sources, mineral agents, and growth factors (vitamins B, amino acids) was examined. The first-order model based on Plackett-Burman design showed that glucose, sodium pyruvate, meat extract and mineral salts significantly influenced the growth of the examined bacteria. The second-order polynomial regression confirmed that maximum biomass production could be achieved by the combination of glucose (12.38 g/l), sodium pyruvate (3.15 g/l), meat extract (4.08 g/l), potassium phosphate (1.46 g/l), sodium acetate (3.65 g/l) and ammonium citrate (1.46 g/l). The validation of the predicted model carried out in bioreactor conditions confirmed the usefulness of the new medium for the culture of L. rhamnosus OXY in large scale. The optimal medium makes the culture of the probiotic bacterium L. rhamnosus OXY more cost effective.

Application of response surface methodology for the optimization of arabinose biotransformation to arabitol by Candida parapsilosis

L-arabitol, a polyol with applications in the food and pharmaceutical industries, is secreted by different yeasts, e.g., Candida spp., Pichia spp., and Debaryomyces spp. The process of its biotechnological production is highly dependent on the physical and chemical conditions of culture. The aim of this study was to use statistical response surface methodology (RSM) to optimize the biotransformation of L-arabinose to arabitol by Candida parapsilosis, a yeast species able to assimilate pentoses. Batch cultures of the yeast were prepared following a Plackett-Burman design for seven variables. Following this, rotation speed, temperature, and L-arabinose concentration were chosen for a central composite design (CCD) experiment, which was carried out to optimize the production L-arabitol. The results showed that the optimal levels for the three factors were: rotation speed 150 rpm, temperature 28°C, and L-arabinose concentration 32.5 g/l. The predicted concentration of arabitol after two days of incubation of C. parapsilosis under the above conditions was 14.3 g/l. The value of R2=0.8323 suggested that this model was well-fitted to the experimental data, and this was confirmed during a verification experiment.

Application of Discriminant Analysis to Distinction of Musical Instruments on the Basis of Selected Sound Parameters

The goal of the presented research was to recognize musical instruments in sound mixes for various levels of accompanying sounds, on the basis of a limited number of sound parameters. Discriminant analysis was used for this purpose. Reduction of the initial large set of sound parameters was performed by means of PCA (principal components analysis), and the factors found using PCA were utilized as input data for discriminant analysis. The results of the discriminant analysis allowed us to assess the accuracy of linear classification on the basis the factors found, and conclude about sound parameters of the highest discriminant power.

Application of Analysis of Variance to Assessment of Influence of Sound Feature Groups on Discrimination between Musical Instruments

In this paper, the influence of the selected sound features on distinguishing between musical instruments is presented. The features were chosen basing on our previous research. Coherent groups of features were created on the basis of significant features, adding complementary ones according to the parameterization method applied, to constitute small, homogenous groups. Next, we investigate (for each feature group separately) if there exist significant differences between means of these features for the studied instruments. We apply multivariate analysis of variance along with post hoc analysis in the form of homogeneous groups, defined by mean values of the investigated features for our instruments. If a statistically significant difference is found, then the homogenous group is established. Such a group may consist of one instrument (distinguished by this feature), or more (instruments similar wrt. this feature). The results show which instruments can be best discerned by which features.

Selection of morphological features of pollen grains for chosen tree taxa

ABSTRACT The basis of aerobiological studies is to monitor airborne pollen concentrations and pollen season timing. This task is performed by appropriately trained staff and is difficult and time consuming. The goal of this research is to select morphological characteristics of grains that are the most discriminative for distinguishing between birch, hazel and alder taxa and are easy to determine automatically from microscope images. This selection is based on the split attributes of the J4.8 classification trees built for different subsets of features. Determining the discriminative features by this method, we provide specific rules for distinguishing between individual taxa, at the same time obtaining a high percentage of correct classification. The most discriminative among the 13 morphological characteristics studied are the following: number of pores, maximum axis, minimum axis, axes difference, maximum oncus width, and number of lateral pores. The classification result of the tree based on this subset is better than the one built on the whole feature set and it is almost 94%. Therefore, selection of attributes before tree building is recommended. The classification results for the features easiest to obtain from the image, i.e. maximum axis, minimum axis, axes difference, and number of lateral pores, are only 2.09 pp lower than those obtained for the complete set, but 3.23 pp lower than the results obtained for the selected most discriminating attributes only. Summary: We present a novel method for selection of the sufficient pollen grains descriptors set, derived from optical microscopic images, for automatic pollen taxa recognition purposes.