Sofia Visa

The Effect of Imbalanced Data Class Distribution on Fuzzy Classifiers - Experimental Study

This study evaluates the robustness of a fuzzy classifier when class distribution of the training set varies. The analysis of the results is based on the classification accuracy and ROC curves. The experimental results reported here show that fuzzy classifiers are less variant with the class distribution and less sensitive to the imbalance factor than decision trees

The control of tomato fruit elongation orchestrated by sun, ovate and fs8.1 in a wild relative of tomato.

Within the cultivated tomato germplasm, sun, ovate and fs8.1 are the three predominant QTLs controlling fruit elongation. Although SUN and OVATE have been cloned, their role in plant growth and development are not well understood. To compare and contrast the effects of the three QTLs in a homogeneous background, we developed near isogenic lines (NILs) in the wild species Solanum pimpinellifolium LA1589 background. We carried out detailed morphological characterization of reproductive and vegetative organs in the single, double and triple NILs and determined the epistatic interactions of the three loci affecting fruit shape. The phenotypic evaluations demonstrated that the three loci regulate unique aspects of ovary and fruit elongation and in different temporal manners. The strongest effect on organ shape was caused by sun. In addition to fruit shape, sun also affected leaf and sepal elongation and stem thickness. The synergistic interaction between sun and ovate or fs8.1 suggested that the pathways involving SUN, OVATE and the gene(s) underlying fs8.1 may converge at a common node. The results of an extensive profiling analysis suggested that the degree of fruit elongation was not related to the accumulation of any of the classical hormones.

Data-driven fuzzy sets for classification

Using the mass assignment mechanism, a fuzzy classifier can be derived directly from the class relative frequency distribution. Moreover, in this framework, a family of fuzzy sets can represent a class, thus adapting the classifier to the need of classification. Graduality and the corresponding concept of error can be used to guide the process of deriving class representing fuzzy sets. The classification algorithm is attractive due to its low complexity. Successful applications include imbalanced data classification problems where the class having fewer examples is the class of interest.

Network Analyses Reveal Shifts in Transcript Profiles and Metabolites That Accompany the Expression of SUN and an Elongated Tomato Fruit

High expression of a gene that controls tomato shape is accompanied by dramatic shifts in gene expression and metabolite and hormone accumulation during the early stages of fruit development. SUN controls elongated tomato (Solanum lycopersicum) shape early in fruit development through changes in cell number along the different axes of growth. The gene encodes a member of the IQ domain family characterized by a calmodulin binding motif. To gain insights into the role of SUN in regulating organ shape, we characterized genome-wide transcriptional changes and metabolite and hormone accumulation after pollination and fertilization in wild-type and SUN fruit tissues. Pericarp, seed/placenta, and columella tissues were collected at 4, 7, and 10 d post anthesis. Pairwise comparisons between SUN and the wild type identified 3,154 significant differentially expressed genes that cluster in distinct gene regulatory networks. Gene regulatory networks that were enriched for cell division, calcium/transport, lipid/hormone, cell wall, secondary metabolism, and patterning processes contributed to profound shifts in gene expression in the different fruit tissues as a consequence of high expression of SUN. Promoter motif searches identified putative cis-elements recognized by known transcription factors and motifs related to mitotic-specific activator sequences. Hormone levels did not change dramatically, but some metabolite levels were significantly altered, namely participants in glycolysis and the tricarboxylic acid cycle. Also, hormone and primary metabolite networks shifted in SUN compared with wild-type fruit. Our findings imply that SUN indirectly leads to changes in gene expression, most strongly those involved in cell division, cell wall, and patterning-related processes. When evaluating global coregulation in SUN fruit, the main node represented genes involved in calcium-regulated processes, suggesting that SUN and its calmodulin binding domain impact fruit shape through calcium signaling.

Rider Transposon Insertion and Phenotypic Change in Tomato

The Rider retrotransposon is ubiquitous in the tomato genome and is likely an autonomous element that still transposes to date. The majority of approximately 2,000 copies of Rider are located near genes. Phenotypes associated with Rider insertion are diverse and often the result of knock out of the underlying genes. One unusual Rider-mediated phenotype resulted from a gene duplication event. By means of read-through transcription, Rider copied part of the surrounding sequence to another location in the genome, leading to high expression of one of the transposed genes, SUN, resulting in an elongated fruit shape. Transcription studies demonstrated that Rider is expressed to levels comparable to the expression of other tomato genes and that control of transposition may be regulated by antisense transcription. Taken together, Rider is a unique retrotransposon that may have played important roles in the evolution of tomato and its closest relatives.

Fuzzy classifiers versus cost-based Bayes classifiers

Learning classifiers for imbalanced data sets is a difficult task for current machine learning algorithms. The difficulty can be traced to the fact that being accuracy driven, most algorithms lead to classifiers which are biased towards the majority class. Introducing in the learning algorithm misclassification costs, which differentiate between classes, has gone a long way towards improving the performance of the resulting classifiers. Alternatively, experiments have shown that a particular type of fuzzy classifiers apply better for imbalanced data sets. This paper explores the hypothesis that fuzzy classifiers can account to a certain extent for the error costs associated with other learning algorithms

A fuzzy classifier for imbalanced and noisy data

This paper deals with the learning concept in the presence of noise (overlap) and imbalance in the training set. The starting assumption is that recognition of the smaller class is much more important than that of the larger class. A fuzzy classifier capable of achieving this based on the relation between fuzzy sets and probability distributions as mediated by the theory of mass assignment. Two approaches to construct fuzzy sets - basic and modified - using the lpd and mpd selection rules are investigated. Preliminary results suggest that the use of mpd selection rule in conjunction with the modified approach is better for recalling the small class at a small cost to the recognition of the negative class.

Modeling of tomato fruits into nine shape categories using elliptic fourier shape modeling and Bayesian classification of contour morphometric data

Classification and characterization of the shape of plant organs are important tools for plant biologists, breeders and growers. Here we use boundary measurements, i.e. contour morphometric data, of scanned tomato fruits in conjunction with elliptic Fourier shape modeling and Bayesian classification techniques to find the optimum number of shape categories. Our findings show that there are nine computationally and visually distinct tomato shape categories: ellipsoid, flat, heart, long, long rectangular, rectangular, round, obovoid, and oxheart. Analyses of fruits from a diverse set of tomato accessions demonstrate that some varieties carry fruits that conform to predominantly one shape category while others carry fruits that conform to multiple shape categories. In particular the categories oxheart and long rectangular feature fruit that tend to equivalently fit several categories of shape, while the flat and obovoid categories contain fruit that consistently conform exclusively to a single category. The findings show that elliptic Fourier shape modeling and Bayesian classification provide an excellent tool for further in depth analyses of fruit shape variation that may occur across varieties and/or result from growth under different environmental conditions.

Obtaining Fuzzy Sets using Mass Assignment Theory - Consistency with Interpolation -

The problem of data summarization as fuzzy sets from frequency distributions is presented. The approach makes use of the mass assignment theory as a framework for unification of fuzzy sets and probability distributions. The approach is consistent with the interpolation of the membership function needed to infer membership degrees for data points previously not seen.

On the Relation between Bayes and Fuzzy Classifiers

A comparison of Bayesian and fuzzy classifiers requires a framework in which a fuzzy set can be related to a probability distribution. This study describes such a framework in which the Bayesian classifier corresponds to a particular choice of the fuzzy classifier. The proposed framework and related concepts is illustrated on a specific classification problem on a well known data set.