Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Classifier generation method using combination of mini-classifiers with regularization and uses thereof

A technology of classifiers and main classifiers, applied in applications, special data processing applications, instruments, etc., can solve problems such as limited performance evaluation and small scale of available sample sets

Inactive Publication Date: 2016-07-06
BIODESIX
View PDF8 Cites 20 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Although previous attempts on this problem have shown efficacy, performance evaluation is limited by cross-validation results due to the small size of the sample set available

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Classifier generation method using combination of mini-classifiers with regularization and uses thereof
  • Classifier generation method using combination of mini-classifiers with regularization and uses thereof
  • Classifier generation method using combination of mini-classifiers with regularization and uses thereof

Examples

Experimental program
Comparison scheme
Effect test

example 1

[0084] Generation of CMC / D classifiers from mass spectrometry

[0085] Data obtained from human samples

[0086] ( Figure 1-Figure 11 )

[0087] This section of the paper will illustrate a practical example of performing the CMC / D classifier development method in the context of a sample set in the form of blood-based samples subjected to mass spectrometry and resulting in a data set used in the classification, The dataset was in the form of 100 features (peaks) at different m / Z positions, which were used as the set of features from which to select the features for the mini-classifier. The above samples were obtained from pancreatic cancer patients participating in a clinical trial of the drug GI-4000. The goal of the classifier generation exercise was to demonstrate whether it is possible to construct a classifier (test) operating on the mass spectrum of a blood-based sample that accurately predicts whether a pancreatic cancer patient associated with that sample is likely ...

example 2

[0176] Classifier generation system and sample test system

[0177] The CMC / D classifier development method described above can be implemented as a tangible classifier development system in the form of a mass spectrometer (or other detection instrument) used to extract samples from multiple samples (e.g., Classifier development set of samples) to obtain mass spectral (or other) data; the general purpose computer has a processing unit that executes code for implementing the CMC / D classification method. In particular, the computer includes a machine-readable memory (such as a hard disk) for storing detection data. The computer also stores executable code that performs preprocessing of the detection data, such as background subtraction, spectral calibration and normalization (as described above), and stores integrated intensity values ​​at specific features for classification, e.g. Example 1 Integrated intensity values ​​for the features listed in Appendix B.

[0178] The compute...

example 3

[0205] EGFR-I Drug Selection (VS2.0) for Non-Small Cell Lung Cancer (NSCLC) Patients Generating CMC / D Classifiers by Mass Spectrometry of Patient Blood-Based Samples

[0206] Another example of generating a CMC / D classifier and using it to guide the treatment of NSCLC patients is described in this section. The generation of the classifier largely followed that described in Example 1 above and in Example 2 above for Figure 11 method described in the Discussion. However, in this example, the processing of the test samples for prediction with the CMC / D classifier utilized the reference spectra, and the spectral Additional adjustments made by processing. The generation of the final classification labels for the tested samples also exploits the feature-dependent noise properties and the following will combine Figure 12 Other techniques are described in more detail. However, this section will show yet another example of generating a CMC / D classifier from mass spectrometry data...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

A method for classifier generation includes a step of obtaining data for classification of a multitude of samples, the data for each of the samples consisting of a multitude of physical measurement feature values and a class label. Individual mini-classifiers are generated using sets of features from the samples. The performance of the mini-classifiers is tested, and those that meet a performance threshold are retained. A master classifier is generated by conducting a regularized ensemble training of the retained / filtered set of mini- classifiers to the classification labels for the samples, e.g., by randomly selecting a small fraction of the filtered mini-classifiers (drop out regularization) and conducting logistical training on such selected mini-classifiers. The set of samples are randomly separated into a test set and a training set. The steps of generating the mini-classifiers, filtering and generating a master classifier are repeated for different realizations of the separation of the set of samples into test and training sets, thereby generating a plurality of master classifiers. A final classifier is defined from one or a combination of more than one of the master classifiers.

Description

[0001] Cross References to Related Applications [0002] This application claims priority under 35 U.S.C § 119(e) to the earlier U.S. Provisional Application US61 / 975,259 filed April 4, 2014, and the earlier U.S. Provisional Application US61 / 878,110 filed September 16, 2013 rights, the contents of these provisional applications are hereby incorporated by reference. technical field [0003] The present disclosure relates to methods and systems for generating classifiers for classifying samples, such as biological samples. The present disclosure features a combination of filtered (filtered) multiple particle-type or "miniature" classifiers that are combined according to a regularized combination method, such as by logically Training and drop-out regularization, and after logistic regression training and drop-out regularization, a master classifier is generated from the above filtered mini-classifiers. [0004] The problem setting for big data challenges in the biological life ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06F19/24G16B40/20
CPCG16B40/00G16B40/20G06V20/698G06F18/285A61B5/7264H01J49/0036H01J49/26
Inventor 海因里希·罗德乔安娜·罗德
Owner BIODESIX
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com