Differences
This shows you the differences between two versions of the page.
Next revision | Previous revision | ||
ai:methods:classification [2013/09/19 16:40] 127.0.0.1 external edit |
ai:methods:classification [2022/04/28 23:27] (current) cyril add missing images |
||
---|---|---|---|
Line 1: | Line 1: | ||
- | ====== Classification ====== | + | ====== Classification ====== |
+ | |||
===== MLP ===== | ===== MLP ===== | ||
+ | |||
Multi Layers Perceptron, //PMC (Perceptron Multi-Couches)// | Multi Layers Perceptron, //PMC (Perceptron Multi-Couches)// | ||
+ | |||
====Gradient Backpropagation==== | ====Gradient Backpropagation==== | ||
+ | |||
// | // | ||
+ | |||
===Stochastic=== | ===Stochastic=== | ||
+ | |||
===with Inertia=== | ===with Inertia=== | ||
+ | |||
===Simulated Annealing=== | ===Simulated Annealing=== | ||
+ | |||
//Recuit Simulé// | //Recuit Simulé// | ||
- | ====newton==== | + | |
- | =Quick Def= | + | |
- | second | + | ====Newton==== |
+ | |||
+ | ==Objective== | ||
+ | |||
+ | Converges faster than gradient descent | ||
+ | |||
+ | ==Quick Def== | ||
+ | |||
+ | Second | ||
+ | |||
=====RBFNN===== | =====RBFNN===== | ||
+ | |||
Radial Basis Functions Neural Networks | Radial Basis Functions Neural Networks | ||
- | * __k-means then gradient | + | |
- | * __incremental | + | ===First method=== |
+ | |||
+ | ==Objective== | ||
+ | |||
+ | You have to chose '' | ||
+ | |||
+ | ==Quick Def== | ||
+ | |||
+ | k-means then gradient | ||
+ | |||
+ | ===Second method=== | ||
+ | |||
+ | ==Quick Def== | ||
+ | |||
+ | incremental | ||
+ | |||
=====SVM===== | =====SVM===== | ||
+ | |||
Support Vectors Machine | Support Vectors Machine | ||
+ | |||
+ | |||
=====Decision tree===== | =====Decision tree===== | ||
- | //arbre de d�cision// | + | |
- | * __ID3__ (based on entropy) | + | //arbre de décision// |
+ | |||
+ | ===ID3=== | ||
+ | |||
+ | ==Quick Def== | ||
+ | |||
+ | based on entropy | ||
+ | |||
=====k-nearest neighbors===== | =====k-nearest neighbors===== | ||
+ | |||
//k plus proches voisins// | //k plus proches voisins// | ||
+ | |||
+ | |||
+ | |||
+ | =====Boosting===== | ||
+ | |||
+ | [Freund, | ||
+ | |||
+ | ==Quick Def== | ||
+ | |||
+ | Consists in combining a lot of weak classifiers to get a strong one. | ||
+ | |||
+ | |||
+ | |||
+ | |||
+ | |||
+ | ====Boosting by majority==== | ||
+ | |||
+ | |||
+ | |||
+ | ====AdaBoost==== | ||
+ | |||
+ | ADAptive BOOSTing, [Freund, | ||
+ | |||
+ | |||
+ | |||
+ | The first and standard version is refered as [[# | ||
+ | |||
+ | |||
+ | |||
+ | ==Quick Def== | ||
+ | |||
+ | Greedy approach | ||
+ | |||
+ | |||
+ | |||
+ | ===Discrete AdaBoost=== | ||
+ | |||
+ | [Freund, | ||
+ | |||
+ | ==References== | ||
+ | |||
+ | - [[|1997, | ||
+ | |||
+ | ==Full Definition== | ||
+ | |||
+ | {{ adaboost-discrete.png | ||
+ | |||
+ | {{ adaboost-discrete_2.png | ||
+ | |||
+ | |||
+ | |||
+ | ===Real AdaBoost=== | ||
+ | |||
+ | [Friedman, | ||
+ | |||
+ | ==References== | ||
+ | |||
+ | - [[|1998, | ||
+ | |||
+ | - [[|1998, | ||
+ | |||
+ | ==Full Definition== | ||
+ | |||
+ | {{ adaboost-real.png | ||
+ | |||
+ | |||
+ | ===LogitBoost=== | ||
+ | |||
+ | [Friedman, | ||
+ | |||
+ | ==References== | ||
+ | |||
+ | - [[|1998, | ||
+ | |||
+ | ==Full Definition== | ||
+ | |||
+ | {{ logitboost.png | ||
+ | {{ logitboost-multi.png | ||
+ | |||
+ | |||
+ | ===Gentle AdaBoost=== | ||
+ | |||
+ | [Friedman, | ||
+ | |||
+ | ==References== | ||
+ | |||
+ | - [[|1998, | ||
+ | |||
+ | ==Full Definition== | ||
+ | |||
+ | {{ adaboost-gentle.png | ||
+ | |||
+ | ===Probabilistic AdaBoost=== | ||
+ | |||
+ | [Friedman, | ||
+ | |||
+ | ==References== | ||
+ | |||
+ | - [[|1998, | ||
+ | |||
+ | |||
+ | ===FloatBoost=== | ||
+ | |||
+ | ==Objective== | ||
+ | |||
+ | AdaBoost is a sequential forward search procedure using the greedy selection strategy to minimize a certain margin on the training set. A crucial heuristic assumption used in such a sequential forward search procedure is the monotonicity (i.e. that addition of a new weak classifier to the current set does not decrease the value of the performance criterion). The premise offered by the sequential procedure in AdaBoost breaks down when this assumption is violated. Floating Search is a sequential feature selection procedure with backtracking, | ||
+ | |||
+ | ==Full Definition== | ||
+ | |||
+ | {{ floatboost.png | ||
+ | |||
+ | |||
+ | |||
+ | ===AdaBoost.Reg=== | ||
+ | |||
+ | [Freund, | ||
+ | |||
+ | ==Objective== | ||
+ | |||
+ | An extension of AdaBoost to regression problems | ||
+ | |||
+ | ==References== | ||
+ | |||
+ | - [[|1997, | ||
+ | |||
+ | | ||
+ | |||
+ | ===Multiclass AdaBoost.M1=== | ||
+ | |||
+ | [Freund, | ||
+ | |||
+ | ==Objective== | ||
+ | |||
+ | Basic extension of AdaBoost to multiclass problems. A weak classifier needs to have an error rate less than 1/2, which is stronger than random guessing (1/k) and is often too difficult to obtain. | ||
+ | |||
+ | ==Quick Def== | ||
+ | |||
+ | A weak classifier associates to an example a label in '' | ||
+ | |||
+ | ==References== | ||
+ | |||
+ | - [[|1997, | ||
+ | |||
+ | ==Full Definition== | ||
+ | |||
+ | {{ adaboost-m1.png | ||
+ | |||
+ | {{ adaboost-m1_2.png | ||
+ | |||
+ | |||
+ | |||
+ | ===Multiclass AdaBoost.M2=== | ||
+ | |||
+ | [Freund, | ||
+ | |||
+ | ==Objective== | ||
+ | |||
+ | Tries to overcome the difficulty of AdaBoost.M1 by extending the communication between the boosting algorithm and the weak learner. The algorithm not only focuses on hard instances, but also on classes which are hard to distinguish. | ||
+ | |||
+ | ==Quick Def== | ||
+ | |||
+ | A weak classifier associates to an example a vector in '' | ||
+ | |||
+ | ==References== | ||
+ | |||
+ | - [[|1997, | ||
+ | |||
+ | ==Full Definition== | ||
+ | |||
+ | {{ adaboost-m2.png | ||
+ | |||
+ | {{ adaboost-m2_2.png | ||
+ | |||
+ | |||
+ | |||
+ | ===Multilabel AdaBoost.MR=== | ||
+ | |||
+ | [Schapire, | ||
+ | |||
+ | ==References== | ||
+ | |||
+ | - [[|1998, | ||
+ | |||
+ | ==Full Definition== | ||
+ | |||
+ | {{ adaboost-r.png | ||
+ | |||
+ | ===Multilabel AdaBoost.MH=== | ||
+ | |||
+ | [Schapire, | ||
+ | |||
+ | ==References== | ||
+ | |||
+ | - [[|1998, | ||
+ | |||
+ | - [[|1998, | ||
+ | |||
+ | ==Full Definition== | ||
+ | |||
+ | {{ adaboost-mh.png | ||
+ | |||
+ | |||
+ | |||
+ | ===Multiclass AdaBoost.MO=== | ||
+ | |||
+ | [Schapire, | ||
+ | |||
+ | ==References== | ||
+ | |||
+ | - [[|1998, | ||
+ | |||
+ | - [[|2006, | ||
+ | |||
+ | |||
+ | |||
+ | ===Multiclass AdaBoost.OC=== | ||
+ | |||
+ | [Schapire, 1997] | ||
+ | |||
+ | ==References== | ||
+ | |||
+ | - | ||
+ | |||
+ | - [[|2006, | ||
+ | |||
+ | |||
+ | |||
+ | ===Multiclass AdaBoost.ECC=== | ||
+ | |||
+ | [Guruswami, | ||
+ | |||
+ | ==References== | ||
+ | |||
+ | - | ||
+ | |||
+ | - [[|2006, | ||
+ | |||
+ | |||
+ | |||
+ | ===AdaBoost.M1W=== | ||
+ | |||
+ | [Eibl, | ||
+ | |||
+ | |||
+ | |||
+ | ===GrPloss=== | ||
+ | |||
+ | [Eibl, | ||
+ | |||
+ | ==References== | ||
+ | |||
+ | - [[|2003, | ||
+ | |||
+ | |||
+ | |||
+ | ===BoostMA=== | ||
+ | |||
+ | [Eibl, | ||
+ | |||
+ | ==References== | ||
+ | |||
+ | - [[|2003, | ||
+ | |||
+ | |||
+ | |||
+ | ===SAMME=== | ||
+ | |||
+ | Stagewise Additive Modeling using a Multi-class Exponential loss function, [Zhu, | ||
+ | |||
+ | |||
+ | |||
+ | ==References== | ||
+ | |||
+ | - [[ |2006, | ||
+ | |||
+ | |||
+ | |||
+ | ===GAMBLE=== | ||
+ | |||
+ | Gentle Adaptive Multiclass Boosting Learning, [Huang, | ||
+ | |||
+ | |||
+ | |||
+ | ==References== | ||
+ | |||
+ | - [[ |2005, | ||
+ | |||
+ | |||
+ | |||
+ | |||
+ | |||
+ | ====UBoost==== | ||
+ | |||
+ | |||
+ | |||
+ | ==Quick Def== | ||
+ | |||
+ | Uneven loss function + greedy | ||
+ | |||
+ | |||
+ | |||
+ | ====LPBoost==== | ||
+ | |||
+ | |||
+ | |||
+ | ==Objective== | ||
+ | |||
+ | Not greedy, exact. | ||
+ | |||
+ | |||
+ | |||
+ | ==References== | ||
+ | |||
+ | - {{2000_Demiriz-Bennett-ShaweTaylor_Linear Programming Boosting via Column Generation.pdf|[Local Copy]}} | ||
+ | |||
+ | |||
+ | |||
+ | ====TotalBoost==== | ||
+ | |||
+ | TOTALly corrective BOOSTing, [Warmuth, | ||
+ | |||
+ | ==References== | ||
+ | |||
+ | - [[|2006, | ||
+ | |||
+ | |||
+ | |||
+ | ====RotBoost==== | ||
+ | |||
+ | [Li, | ||
+ | |||
+ | |||
+ | |||
+ | ==References== | ||
+ | |||
+ | - {{2003_Li-AbuMostafa-Pratap_CGBoost - Conjugate Gradient in Function Space.pdf|[Local Copy]}} | ||
+ | |||
+ | |||
+ | |||
+ | ====alphaBoost==== | ||
+ | |||
+ | [Li, | ||
+ | |||
+ | |||
+ | |||
+ | ==References== | ||
+ | - {{2003_Li-AbuMostafa-Pratap_CGBoost - Conjugate Gradient in Function Space.pdf|[Local Copy]}} | ||
+ | |||
+ | |||
+ | ====MILBoost==== | ||
+ | (Multiple Instance Learning BOOSting), [Viola, | ||
+ | |||
+ | ==References== | ||
+ | - {{|2005, | ||
+ | |||
+ | |||
+ | ====CGBoost==== | ||
+ | |||
+ | Conjugate Gradient BOOSTing, [Li, | ||
+ | |||
+ | |||
+ | ==References== | ||
+ | - {{2003_Li-AbuMostafa-Pratap_CGBoost - Conjugate Gradient in Function Space.pdf|[Local Copy]}} | ||
+ | |||
+ | |||
+ | |||
+ | |||
+ | |||
+ | ====Bootstrap Aggregating==== | ||
+ | |||
+ | |||
+ | |||
+ | =====Cascades of detectors===== | ||
+ | |||
+ | ==Quick Def== | ||
+ | |||
+ | A cascade of classifiers is a degenerated decision tree, where at each stage a classifier is trained to detect almost all objects of interest, while rejecting a certain fraction of the non-object patterns (eg eliminates 50% of non-object patterns and falsely eliminates 0.1%, then after 20 stages it can be expected a false alarm rate of 0.5^20 and a hit rate of 0.999^20). It enables to focus attention on certain regions and dramatically increases speed. | ||
+ | |||
+ | |||
+ | |||
+ | =====Trees of detectors===== | ||
+ | |||
+ | |||
====== Regression ====== | ====== Regression ====== | ||
+ | |||
+ | |||
* **MLP (Multi Layers Perceptron)** | * **MLP (Multi Layers Perceptron)** | ||
+ | |||
* **RBFNN (Radial Basis Functions Neural Network)** | * **RBFNN (Radial Basis Functions Neural Network)** | ||
+ | |||
* **SVR (Support Vectors Regressor)** | * **SVR (Support Vectors Regressor)** | ||
+ | |||