Support vector machine (SVM) is a learning method developed in the mid-1990s based on statistical theory. SVM classifier currently more popular classifier. This paper presents boundary detection technique for retaining potential support vector. Through seeking to structural risk minimization of SVM, it improves generalization ability and achieves empirical confidence range case small sample size can also obtain desired good law.

To improve the prediction effect of time series, we make a systematic study various series methods based on statistics and machine learning in this paper. In experiment, compare results several methods. particular, much research has been done selection experimental data because representative can better test effectiveness practicability method. Based idea divide conquer complex problems strategy continuous optimization learning, proposed LSTM-TFE, LR-TFE, BR-TFE combined EEMD, LSTM, LR, BR...

In recent years, with the rapid development of wind power generation, some problems are gradually highlighted.At present, one essential methods to solve these is predict speed.In this paper, a hybrid BRR-EEMD method proposed for short-term speed prediction based on Bayesian ridge regression and ensemble empirical mode decomposition.We use decomposition decompose complex time series into several relatively milder, more regular, stable subsequences.Then each subsequence carried out by using...

The stock market is a chaotic, complex, and dynamic financial market. prediction of future prices concern controversial research issue for researchers. More more analysis methods are proposed by We hybrid method the using LSTM ensemble EMD in this paper. use comprehensive to decompose complex original price time series into several subsequences which smoother, regular stable than series. Then, we train predict each subsequence. Finally, obtained values fused subsequences. In experiment,...

In general, the stock trend is mainly driven by big order transactions. Believing that rise with a large volume closely associated net inflow, we propose an efficient recommendation model based on inflow in paper. to compute of stock, use M/G/1 queue system measure all tick-by-tick transaction data. Based indicator select some stocks higher value constitute prerecommended set for target investor user. recommend which this style familiar them users, divide lots investors into several...

This paper introduces a multiscale multifractal multiproperty analysis based on Rényi entropy (3MPAR) method to analyze short-range and long-range characteristics of financial time series, then applies this the five series properties in four stock indices. Combining two techniques detrended fluctuation (MFDFA), 3MPAR focuses curves generalized Hurst exponent which allows us study more universal subtle series. By analyzing profiles logarithm distribution MFDFA indices, shows some markets....

Information hiding technology is a hot spot in information security, and applied the fields of digital multimedia copyright protection secret communication. According to analysis characteristics browser parsing HTML web page little capacity available for hided page, new efficient method with tag attributes has been proposed this paper, which overcomes shortcoming ability imperceptibility contradict machine filtration traditional algorithms improves embedded other algorithm based on...

The support vector machine (SVM) is a learning method developed based on statistical theory. SVM widely used in classification and prediction. Since the financial time series complex, traditional forecasting methods are less reliable. In this paper, we research machine. Although speed of prediction process slow, it can improve accuracy series. experimental results show approach

To effectively improve the power dispatching, prediction accuracy of wind has been concern many scholars for years. The problem is actually equivalent to speed problem. Based on linear regression (LR) and variational mode decomposition (VMD), in this paper, we proposed an efficient hybrid method predict speed. In method, VMD used decompose signal into several sub-signal. Compared with original wind-speed series, each sub-signal a more stable subsequence signal. Then, LR Eventually, obtain...

Abstract Accurate skin lesion segmentation (SLS) plays an essential role in the computer‐aided diagnosis of diseases, for example, melanomas. However, automated SLS is challenging due to variations nature particularly ambiguities boundaries areas (SLAs) and occlusions by hairs. Though many deep learning models, represented UNet, have been successfully applied over past years, most suffer from inaccurate SLA with heavy model parameters slow inferencing speed, hindering their practical...

The data of crop diseases and insect pests have the characteristics massive, heterogeneous multi-source. traditional storage methods systems problems low efficiency weak scalability. To solve these problems, this paper introduces a HBase based system mining technology early warning on Hadoop. can store manage pest persistently, make up for deficiency relational database, improve speed query warning, provide new method in big environment.

This paper introduces a multiscale multifractal method based on n-mean to analyze the and role of n-day mean individual share exchange or stock composite index data. The also allows us discuss financial time series depending their magnitude scale using generalized Hurst exponent surfaces exponent. In this paper, proposed focuses data, which lets apply MMA all series. From experimental results, we find that results provide important basic information for study It is necessary

Combining with enterprises and implementing the integration of industry education has become a common practice in professional construction talent training many local undergraduate colleges. However, under background engineering certification, how to effectively implement high-quality is problem worthy exploration. Starting from systematicness training, combined elements certification integration, this paper puts forward computer mode driven by analyzes specific implementation strategy...

This article first briefly introduces the basic principles of multipath effects and analyzes causes fading, then gives standards definition for some related concepts rigorous mathematical expressions those. Then, based on a theoretical model, using dynamic system modeling simulation software package Simulink Matlab software, this corresponding vividly recreates actual process fading channels. The results prove feasibility algorithm have good performances higher applied value.

In the conventional coding techniques, all follow Nyquist Law. The law states that sampling rate is higher than twice frequency of original signal. For this method can not overcome enormous amount computation and a waste resources, compressed sensing theory recently proposed for image compression, greatly reduce rate, paper focuses on compression algorithm based theory, simulation results prove feasibility has good performances in invisibility applied value.