scispace - formally typeset
Search or ask a question
Topic

AdaBoost

About: AdaBoost is a research topic. Over the lifetime, 6254 publications have been published within this topic receiving 257741 citations.


Papers
More filters
Patent
22 Jun 2018
TL;DR: In this article, a smart discount strategy providing system applied to bills is presented, which includes a data collection module, a data storage module, data backup module, and a data safety management module.
Abstract: The invention discloses a smart discount strategy providing system applied to bills. The system includes a data collection module, a data storage module, a data backup module, a data safety managementmodule and a bill discount strategy module. According to the invention, through machine learning, algorithm calculation and automatic investigation and survey of market data and user behavior preference, and according to bill payment timely level of clients, discount strategies of different amplitudes and different application ranges are made. At the same time, an Adaboost is used as a classifierwith high classification precision, the discount strategies having difference are generated and different discount levels can be provided to different clients. In an Adaboost frame, different regression classification models are used for constructing a weak learning machine; high flexibility is achieved. The flexible construction system is convenient for customization and development. When the system acts as a simple binary classifier, a simple structure is realized and intelligible results are provided, and overfitting is prevented.
Proceedings ArticleDOI
01 Oct 2019
TL;DR: In this paper, a face recognition method based on PCA and Back Propagation Neural Network is presented. But the method uses the Adaboost learning algorithm to select a small number of important feature from a large data set.
Abstract: This paper presented a face recognition method based on PCA and Back Propagation Neural Network which is robust and simple algorithm based on image captured by camera. The proposed algorithm is done on three stages, namely face detection, face feature extraction and face recognition. The face detection is performed using Haar-Like Feature. The Haar-Like feature analyzes the pixels in the image into squares by function. The method uses the Adaboost learning algorithm to select a small number of important feature from a large data set. The detected face is then extracted using PCA. PCA will select and reduce the face feature based on eigenvalues of correlation matrix data. We obtain 500 features face extracted using PCA. The face feature becomes the input to neural network for recognition process. The neural network is developed with two hidden layers with 15 nodes on each and three nodes of output layer. The experiment is performed in real time environment. Using 5 faces image data with each data is taken 100 times, the experimental result showed the satisfactory result with 87.5% recognition rate in average.
Journal ArticleDOI
TL;DR: This paper puts forward a visual method of detecting pop-cans based on AdaBoost Algorithm, and it shows that these pop-can detectors trained by this visual method have high detection rate and speed.
Abstract: This paper puts forward a visual method of detecting pop-cans based on AdaBoost Algorithm. This method, is basing on the idea of AdaBoost Algorithm, and we use Haar Feature and LBP Feature to extract pop-can characteristics respectively. Finally we compare the differences of the training processes and the experiment results between these two detectors. It shows that these pop-can detectors that are trained by this visual method have high detection rate and speed.
Journal Article
TL;DR: A camera based assistive text reading framework to help blind persons read text labels and product packaging from hand-held object in their daily resides and a novel text localization algorithm by learning gradient features of stroke orientations and distributions of edge pixels in an Adaboost model.
Abstract: A camera based assistive text reading framework to help blind persons read text labels and product packaging from hand-held object in their daily resides is proposed. To isolate the object from cluttered backgrounds or other surroundings objects in the camera view, we propose an efficient and effective motion based method to define a region of interest (ROI) in the video by asking the user to shake the object. In the extracted ROI, text localization and recognition are conducted to acquire text information. To automatically localize the text regions from the object ROI, we propose a novel text localization algorithm by learning gradient features of stroke orientations and distributions of edge pixels in an Adaboost model. Text characters in the localized text regions are then binarized and recognized by off-the shelf optical character recognition software. The recognized text codes are output to blind users in speech.
Proceedings ArticleDOI
03 Mar 2013
TL;DR: This paper examines the MPDT on a random selection of 26 benchmark data sets from the UCI repository and compared it with Bagging, AdaBoost and C4.5.
Abstract: Decision trees are well-known and established models for classification and regression. In this paper, we propose multi-path decision tree algorithm (MPDT). Different from traditional decision tree where the path of each record is deterministic and exclusive, a record can trace several paths simultaneously in multi-path decision tree so that it has the effect of ensemble classifiers with only one classifier. Local class information gain is the value of class information (entropy or Gini, etc) given the value of an attribute relative to class information unsupervised. We examine the MPDT on a random selection of 26 benchmark data sets from the UCI repository and compared it with Bagging, AdaBoost and C4.5. The results note that MPDT has better performance. Keywords-component; formatting; style; styling; insert

Network Information
Related Topics (5)
Feature extraction
111.8K papers, 2.1M citations
93% related
Feature (computer vision)
128.2K papers, 1.7M citations
92% related
Deep learning
79.8K papers, 2.1M citations
92% related
Convolutional neural network
74.7K papers, 2M citations
92% related
Artificial neural network
207K papers, 4.5M citations
92% related
Performance
Metrics
No. of papers in the topic in previous years
YearPapers
2023490
20221,088
2021473
2020410
2019436
2018369