Robust Support Vector Machines with Polyhedral Uncertainty of the Input Data

Research every college will continue to grow. Research will be stored in softcopy and hardcopy. The preparation of the research should be categorized in order to facilitate the search for people who need reference. To categorize the research, we need a method for text mining, one of them is with the implementation of Support Vector Machines (SVM). The data used to recognize the characteristics of each category then it takes secondary data which is a collection of abstracts of research. The data will be pre-processed with several stages: case folding converts all the letters into lowercase, stop words removal removal of very common words, tokenizing discard punctuation, and stemming searching for root words by removing the prefix and suffix. Further data that has undergone preprocessing will be converted into a numerical form with for the term weighting stage that is the weighting contribution of each word. From the results of term weighting then obtained data that can be used for data training and test data. The training process is done by providing input in the form of text data that is known to the class or category. Then by using the Support Vector Machines algorithm, the input data is transformed into a rule, function, or knowledge model that can be used in the prediction process. From the results of this study obtained that the categorization of research produced by SVM has been very good. This is proven by the results of the test which resulted in an accuracy of 90%.

Download Full-text

A Novel Approach for the Detection, Classification and Localization of Transmission Lines Faults using Wavelet Transform and Support Vector Machines Classifier

International Journal of Engineering & Technology ◽

10.14419/ijet.v7i2.17.11559 ◽

2018 ◽

Vol 7 (2.17) ◽

pp. 56

Author(s):

Nilish S.Wani ◽

Dr R. P. Singh

Keyword(s):

Artificial Intelligence ◽

Wavelet Transform ◽

Support Vector Machines ◽

Transmission Lines ◽

Input Data ◽

Test System ◽

Support Vector ◽

Discrete Wavelet ◽

Output Data ◽

Vector Machines

Over the time, techniques have been developed whose approach has been linked mainly to the diagnosis of faults, which is why over the years these techniques have been improved little by little in order to complement or at best cases innovate the traditional methods used for the detection of faults from the mathematical point of view relying mainly on sophisticated methods and some of them related to artificial intelligence. Taking into account the aforementioned, in this paper we propose the use one of the branches of artificial intelligence, specifically automatic learning through the tool known as Support Vector Machines (SVM) to find a method with which is feasible to identify and classify the type of fault. For the creation of the mathematical model it is essential to have a database. The database consists of input data and output data, the input data are the detail coefficients obtained from the decomposition of the current and voltage signals using the Discrete Wavelet Transform (DWT). Meanwhile, the output data are the labels assigned and with which the model can identify and classify the different types of faults. Both current signals and voltage signals are generated based on an extensive simulation of faults along the longest transmission line that has a test system.

Download Full-text

Predicting the strength of Populus spp. clones using artificial neural networks and ε-regression support vector machines (ε-rSVM)

Holzforschung ◽

10.1515/hf.2011.107 ◽

2011 ◽

Vol 65 (6) ◽

pp. 855-863 ◽

Cited By ~ 10

Author(s):

Shawn D. Mansfield ◽

Kyu-Young Kang ◽

Lazaros Iliadis ◽

Stavros Tachos ◽

Stavros Avramidis

Keyword(s):

Neural Networks ◽

Artificial Neural Networks ◽

Support Vector Machines ◽

Input Data ◽

Back Propagation ◽

Gaussian Kernel ◽

Support Vector ◽

Data Set ◽

Vector Machines ◽

Artificial Neural

Abstract Wood properties, including bending stiffness and strength, basic density and microfibril angle were experimentally obtained for six aspen and six hybrid poplar clones grown in Western Canada. Data analysis attempted to establish a relationship between wood mechanical properties and intrinsic wood attributes by means of artificial neural networks (ANN) and ε-regression support vector machines (ε-rSVM) employing a 5-fold cross validation approach (5-fold CV). Initial results for strength were acceptable, but require further improvement. Estimations of stiffness results (MOE) were inferior to those of strength (MOR) due to the fact that in several regression cases, the developed model worked well for narrow windows of data, but failed on a large scale due to the high variations in the values of the input data vectors. In such cases, the result is probably the development of regression with uneven performance throughout the input data set, and therefore the modeling capacity is poor. To avoid this predicament, different neural networks with one output neuron were developed in order to estimate either the stiffness or the strength, and at the same time the approximation capabilities of ε-rSVM were employed. In both methods, 5-fold CV was carried out in order to attain a more generalized solution by eliminating the boundary effect phenomena and by avoiding local behavior of the global support vector regression. The resultant models were evaluated by common metrics. The best ANN for the estimation of strength in combination with 5-fold CV, was a modular back propagation with average R2=0.70, and mean root mean square error (MRMSE) equal to 0.19 and mean average percent error (MAPE) equal to 12.5%. The Gaussian kernel 5-fold CV ε-rSVM estimated MOR with similar accuracy. The best 5-fold CV ANN for MOE estimation was a feed forward back propagation one, with average R2=0.60, MRMSE equal to 0.23 and MAPE equal to 41.5%, which was better than all other kernel methods employed.

Download Full-text