Multi-class classifiers often compute scores for the classification samples describing probabilities to belong to different classes. In order to improve the performance of such classifiers, machine learning experts need to analyze classification results for a large number of labeled samples to find possible reasons for incorrect classification. Confusion matrices are widely used for this purpose. However, they provide no information about classification scores and features computed for the samples. We propose a set of integrated visual methods for analyzing the performance of probabilistic classifiers. Our methods provide insight into different aspects of the classification results for a large number of samples. One visualization emphasizes at which probabilities these samples were classified and how these probabilities correlate with classification error in terms of false positives and false negatives. Another view emphasizes the features of these samples and ranks them by their separation power between selected true and false classifications. We demonstrate the insight gained using our technique in a benchmarking classification dataset, and show how it enables improving classification performance by interactively defining and evaluating post-classification rules.

Download full-text PDF

Source
http://dx.doi.org/10.1109/TVCG.2014.2346660DOI Listing

Publication Analysis

Top Keywords

classification
9
visual methods
8
methods analyzing
8
classification large
8
large number
8
samples
6
analyzing probabilistic
4
probabilistic classification
4
classification data
4
data multi-class
4

Similar Publications

Although empirical support for the International Statistical Classification of Diseases and Related Health Problems (11th ed.; ICD-11) distinction between posttraumatic stress disorder (PTSD) and complex PTSD (CPTSD) is growing, research into the ICD-11 CPTSD model in prison staff is lacking. This study used latent profile analysis (LPA) to (a) determine if there are distinct groups of trauma-exposed prison governors (i.

View Article and Find Full Text PDF

Background And Objectives: Low-birth weight, premature infants often have severe intraventricular hemorrhage (IVH), which can result in posthemorrhagic hydrocephalus (PHH), sometimes requiring cerebrospinal fluid diversion. Initial temporizing management of PHH includes placement of a ventriculosubgaleal shunt (VSGS) or ventricular access device (VAD). Studies have found similar permanent shunt conversion rates between VSGS and VAD but were limited by sample scope and size.

View Article and Find Full Text PDF

Studies on trematodes and acanthocephalans from freshwater fishes of Hubei Province, central China, with the erection of a new genus Quadrihexaspiron gen. n. (Acanthocephala: Neoechinorhynchidae).

Folia Parasitol (Praha)

January 2025

State Key Laboratory of Freshwater Ecology and Biotechnology, and Laboratory of Fish Diseases, Institute of Hydrobiology, Chinese Academy of Sciences, Wuhan 430072, Hubei Province, People's Republic of China *Address for correspondence: Frantisek Moravec, Institute of Parasitology, Biology Centre of the Czech Academy of Sciences, Branisovska 31, 370 05 Ceske Budejovice, Czech Republic. E-mail: ORCID: 0000-0003-1086-1181.

The present paper comprises a systematic survey of trematodes and acanthocephalans based on helminthological examinations of 64 specimens of 14 species of freshwater fishes, belonging to six families of four fish orders, mostly from localities in Hubei Province, central China, collected in the autumn of 2002. A total of 15 trematode species (in 12 families) and 5 acanthocephalan species (in four families) was recorded. Almost all parasites are briefly described and illustrated and problems concerning their morphology, taxonomy, hosts and geographical distribution are discussed.

View Article and Find Full Text PDF

MultiTax-human: an extensive and high-resolution human-related full-length 16S rRNA reference database and taxonomy.

Microbiol Spectr

January 2025

State Key Laboratory for Diagnosis and Treatment of Infectious Diseases, National Clinical Research Center for Infectious Diseases, National Medical Center for Infectious Diseases, Collaborative Innovation Center for Diagnosis and Treatment of Infectious Diseases, The First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, China.

Considering that the human microbiota plays a critical role in health and disease, an accurate and high-resolution taxonomic classification is thus essential for meaningful microbiome analysis. In this study, we developed an automatic system, named MultiTax pipeline, for generating taxonomy from full-length 16S rRNA sequences using the Genome Taxonomy Database and other existing reference databases. We first constructed the MultiTax-human database, a high-resolution resource specifically designed for human microbiome research and clinical applications.

View Article and Find Full Text PDF

Nuclear Factor Y (NF-Y) represents a group of transcription factors commonly present in higher eukaryotes, typically consisting of three subunits: NF-YA, NF-YB, and NF-YC. They play crucial roles in the embryonic development, photosynthesis, flowering, abiotic stress responses, and other essential processes in plants. To better understand the genome-wide NF-Y domain-containing proteins, the protein physicochemical properties, chromosomal localization, synteny, phylogenetic relationships, genomic structure, promoter -elements, and protein interaction network of NtNF-Ys in tobacco ( L.

View Article and Find Full Text PDF

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!