Statistical learning theory

About: Statistical learning theory is a research topic. Over the lifetime, 1618 publications have been published within this topic receiving 158033 citations.

Parallel support vector machines for multi-category classification of large scale data

Abstract: of a Dissertation Submitted to the Graduate Studies Office of The University of Southern Mississippi in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

Learning from Data - A SVM Based Approach

Abstract: Support Vector Machines were introduced by Vapnik within the area of statistical learning theory and structural risk minimization and first applied to classification problems as alternatives to multi-layer neural networks. The high generalization ability provided by Support Vector Classifiers has inspired recent work on computational speedups as well as the fundamental theory of model complexity and generalization. In this paper we present a gradient-type learning algorithm for training a SVM and analyze its performance when applied to linear separable data. In order to evaluate the efficiency of our learning method, several tests were performed, the conclusions being formulated in the final section of the paper.

A Sparse Data Preprocessing Using Support Vector Regression

Abstract: In various fields as web mining, bioinformatics, statistical data analysis, and so forth, very diversely missing values are found. These values make training data to be sparse. Largely, the missing values are replaced by predicted values using mean and mode. We can used the advanced missing value imputation methods as conditional mean, tree method, and Markov Chain Monte Carlo algorithm. But general imputation models have the property that their predictive accuracy is decreased according to increase the ratio of missing in training data. Moreover the number of available imputations is limited by increasing missing ratio. To settle this problem, we proposed statistical learning theory to preprocess for missing values. Our statistical learning theory is the support vector regression by Vapnik. The proposed method can be applied to sparsely training data. We verified the performance of our model using the data sets from UCI machine learning repository.

Izrada strategija trgovanja električnom energijom u Njemačkoj korištenjem metode potpornih vektora

Abstract: Machine learning is the theory of programming computers so that they optimize certain success criteria regarding data examples or previous experience [7]. The most important theorem for machine learning is the PAC theorem that says: if a model does well enough on most examples, it is probably good enough. Machine learning model is trained on training examples, in which it is tried, to achieve optimal adjustment. An attempt is made to avoid underfitting, (when the model does poorly on training examples), and overfitting, (when the model does perfectly on training examples, but makes mistakes on new examples). Support Vectors Machine was created by Vladimir Vapnih in 1960’s, but was not used until 1992, when Isabelle Guyon, Bernhard Boser and Vladimir Vapnik introduced nonlinear classifier using the kernel trick. Regression version was proposed in 1997, by Vladimir Vapnik, Harris Drucker, Chris Burges, Linda Kaufman and Alex Smola. Statistical learning theory seeks to minimize the risk and complexity of the model. Support vectors machine is based on statistical learning theory. In support vectors machine, the minimization of the risk and complexity of the model is aimed by finding optimal separating hyper plane. Optimal hyper plane is the one with smallest distance between hyper plane and support vectors. In situations when data is not separable or time to discover optimal hyper plane is too long, kernel trick is used: points are mapped to higher dimensional space than original. The marketplace is an organized and centralized place of trading, where sellers and buyers present their offers. Lines of supply and demand are created considering all received offers. Market price and quantity for trading, are determined upon intersection of the supply and demand line. Trading strategy begins by seeking a model with the most precision in predicting market price of electric energy, and it differs regarding trader role on the market. This thesis elaborates the example of the strategy development of electricity trading in Germany by using the support vectors machine. Data used are: temperature, production of electric energy from solar power plants, production of electric energy from wind power plants, production of electric energy from conventional power plants over 100 MW output power, production of electric energy from hydro power plants, energy of precipitation, and price of electric energy measured in hourly time intervals. All data have variations with extreme jumps between values. Annually, weekly and daily patterns are shown in the price of electric energy and temperature. Similar patterns are seen in the production of electric energy from solar power plants, wind power plants, and from conventional power plants. There is a distinct connection between the production of electric energy from hydro power plants, and the energy of precipitation. Functions and classes of programming language Python were used for data preparation and algorithm implementation. Data from 7. 5. to 11. 11. 2014 were used for training of model, and data from 12. 11. to 19. 11. 2014 were used for model validation. Model with selection, \(C=1400\) and \(\gamma=9.9999999999999995 e^{-03}\) gave the best results with mean absolute percentage error \(14.14 \% \). It is possible to create a strategy of electricity trading based on data produces by described model. Example of one such strategy is given in this thesis for 12. 11. 2014.

Analysis of Economic Risk Identification Evaluation and Management and Control Strategy Based on Support Vector Machine Algorithm

Abstract: With the deepening of economic globalization, China's participation in global foreign trade activities is constantly diversified, and at the same time it faces more and more risks. China's long-term accumulated risks in the future may be released intensively, causing high incidence. Potential risks in macro-economy, business environment, sovereign credit, debt and other fields will bring certain losses to enterprises and even hinder their development. Support Vector Machine (SVM) based on statistical learning theory is a new machine learning algorithm, which can successfully deal with classification and regression problems. Because of the excellent learning performance of SVM, this technology has become a research hotspot in current academic circles. This paper expounds the basic theory of support vector machine in detail, constructs the basic framework of economic investment risk prediction model based on support vector machine, and gives the concrete steps to realize the model and the key problems to be solved. The algorithm is very practical. Using support vector machine regression theory to predict economic risk not only enriches the existing risk prediction methods in theory, but also has important value in practical application.