由于特征提取是数据挖掘的基础工作,而其质量对挖掘结果有很大影响,为此针对局部线性嵌入(LLE:Locally Linear Embedding)算法并未考虑同一数据的不同特征之间的相关性,不能较好地保留时间信号的主要形态趋势,提出了基于特征相关性的局...由于特征提取是数据挖掘的基础工作,而其质量对挖掘结果有很大影响,为此针对局部线性嵌入(LLE:Locally Linear Embedding)算法并未考虑同一数据的不同特征之间的相关性,不能较好地保留时间信号的主要形态趋势,提出了基于特征相关性的局部线性嵌入(CC-LLE:Local Linear Embedding Algorithm Based on Characteristic Correlation)算法,并应用于轴承故障诊断。针对轴承故障信号周期性特点,该算法在特征提取阶段对数据进行分段操作,选取各分段上的标准偏差作为特征,构造原始数据的特征样本集,从而有效提取鉴别特征。通过在轴承数据集上进行实验验证了该算法在特征提取方面的有效性。展开更多
A new prediction technique is proposed for chaotic time series. The usefulness of the technique is that it can kick off some false neighbor points which are not suitable for the local estimation of the dynamics sys...A new prediction technique is proposed for chaotic time series. The usefulness of the technique is that it can kick off some false neighbor points which are not suitable for the local estimation of the dynamics systems. A time-delayed embedding is used to reconstruct the underlying attractor, and the prediction model is based on the time evolution of the topological neighboring in the phase space. We use a feedforward neural network to approximate the local dominant Lyapunov exponent, and choose the spatial neighbors by the Lyapunov exponent. The model is tested for the Mackey-Glass equation and the convection amplitude of lorenz systems. The results indicate that this prediction technique can improve the prediction of chaotic time series.展开更多
The paper＇s aim is how to forecast data with variations involving at times series data to get the best forecasting model. When researchers are going to forecast data with variations involving at times series data （i...The paper＇s aim is how to forecast data with variations involving at times series data to get the best forecasting model. When researchers are going to forecast data with variations involving at times series data （i.e., secular trends, cyclical variations, seasonal effects, and stochastic variations）, they believe the best forecasting model is the one which realistically considers the underlying causal factors in a situational relationship and therefore has the best ＂track records＂ in generating data. Paper＇s models can be adjusted for variations in related a time series which processes a great deal of randomness, to improve the accuracy of the financial forecasts. Because of Na＇fve forecasting models are based on an extrapolation of past values for future. These models may be adjusted for seasonal, secular, and cyclical trends in related data. When a data series processes a great deal of randomness, smoothing techniques, such as moving averages and exponential smoothing, may improve the accuracy of the financial forecasts. But neither Na＇fve models nor smoothing techniques are capable of identifying major future changes in the direction of a situational data series. Hereby, nonlinear techniques, like direct and sequential search approaches, overcome those shortcomings can be used. The methodology which we have used is based on inferential analysis. To build the models to identify the major future changes in the direction of a situational data series, a comparative model building is applied. Hereby, the paper suggests using some of the nonlinear techniques, like direct and sequential search approaches, to reduce the technical shortcomings. The final result of the paper is to manipulate, to prepare, and to integrate heuristic non-linear searching methods to serve calculating adjusted factors to produce the best forecast data.展开更多
Analyzing time series characteristics of red tide is the basis of disaster prevention and mitigation,which is very important to red tide prediction.There are trend comp onents and periodic components in annual time se...Analyzing time series characteristics of red tide is the basis of disaster prevention and mitigation,which is very important to red tide prediction.There are trend comp onents and periodic components in annual time series of occurrence freque ncy and area of red tides,so Gray-Periodic Extensional Combinatorial Model(GPECM)is used to extract these components.The fitting degree of occurrence frequency and area can reach 95.20% and 95.24%,respectively.The performance of GPECM is better than Gray Model,Fourier Series Extension Model,and Holt-Winter Exponential Smoothing Model in model stability.Consequently,it is used to forecast the occurrence frequency and area in 2020 and 2021,and results show that the annual frequency of red tides in 2020 and 2021 can rise to 39 and 41,respectively,and that the annual occurrence area of red tides can rise to 3168 km^(2),which is about 59% more than last year.In 2021,it can fall to 1901 km^(2).展开更多
Noise can induce inverse period-doubling transition and chaos. The effects of the colored noise on periodic orbits, of the different periodic sequences in the logistic map, are investigated. It is found that the dynam...Noise can induce inverse period-doubling transition and chaos. The effects of the colored noise on periodic orbits, of the different periodic sequences in the logistic map, are investigated. It is found that the dynamical behaviors of the orbits, induced by an exponentially correlated colored noise, are different in the mergence of transition, and the effects of the noise intensity on their dynamical behaviors are different from the effects of the correlation time of noise. Remarkably, the noise can induce new periodic orbits, namely, two new orbits emerge in the period-four sequence at the bifurcation parameter value μ = 3.5, four new orbits in the period-eight sequence at μ= 3.55, and three new orbits in the period-six sequence at μ = 3.846, respectively. Moreover, the dynamical behaviors of the new orbits clearly show the resonancelike response to the colored noise.展开更多
A method of the fuzzy cross-correlation factor exponent in dynamics is researched and proposed to diagnose abnormality of cracks in the concrete dam. Moreover, the Logistic time series changing from period-doubling bi...A method of the fuzzy cross-correlation factor exponent in dynamics is researched and proposed to diagnose abnormality of cracks in the concrete dam. Moreover, the Logistic time series changing from period-doubling bifurcation to chaos is tested first using this method. Results indicate that it can distinguish inherent dynamics of time series and can detect mutations. Considering that cracks in the concrete dam constitute an open, dissipative and complex nonlinear dynamical system, a typical crack on the downstream face of a concrete gravity arch dam is analyzed with the proposed method. Two distinct mutations are discovered to indicate that the abnormality diagnosis of cracks in the concrete dam is achieved dynamically through this method. Furthermore, because it can be directly utilized in the measured crack opening displacement series to complete abnormality diagnosis, it has a good prospect for practical applications.展开更多
文摘由于特征提取是数据挖掘的基础工作,而其质量对挖掘结果有很大影响,为此针对局部线性嵌入(LLE:Locally Linear Embedding)算法并未考虑同一数据的不同特征之间的相关性,不能较好地保留时间信号的主要形态趋势,提出了基于特征相关性的局部线性嵌入(CC-LLE:Local Linear Embedding Algorithm Based on Characteristic Correlation)算法,并应用于轴承故障诊断。针对轴承故障信号周期性特点,该算法在特征提取阶段对数据进行分段操作,选取各分段上的标准偏差作为特征,构造原始数据的特征样本集,从而有效提取鉴别特征。通过在轴承数据集上进行实验验证了该算法在特征提取方面的有效性。
文摘A new prediction technique is proposed for chaotic time series. The usefulness of the technique is that it can kick off some false neighbor points which are not suitable for the local estimation of the dynamics systems. A time-delayed embedding is used to reconstruct the underlying attractor, and the prediction model is based on the time evolution of the topological neighboring in the phase space. We use a feedforward neural network to approximate the local dominant Lyapunov exponent, and choose the spatial neighbors by the Lyapunov exponent. The model is tested for the Mackey-Glass equation and the convection amplitude of lorenz systems. The results indicate that this prediction technique can improve the prediction of chaotic time series.
文摘The paper＇s aim is how to forecast data with variations involving at times series data to get the best forecasting model. When researchers are going to forecast data with variations involving at times series data （i.e., secular trends, cyclical variations, seasonal effects, and stochastic variations）, they believe the best forecasting model is the one which realistically considers the underlying causal factors in a situational relationship and therefore has the best ＂track records＂ in generating data. Paper＇s models can be adjusted for variations in related a time series which processes a great deal of randomness, to improve the accuracy of the financial forecasts. Because of Na＇fve forecasting models are based on an extrapolation of past values for future. These models may be adjusted for seasonal, secular, and cyclical trends in related data. When a data series processes a great deal of randomness, smoothing techniques, such as moving averages and exponential smoothing, may improve the accuracy of the financial forecasts. But neither Na＇fve models nor smoothing techniques are capable of identifying major future changes in the direction of a situational data series. Hereby, nonlinear techniques, like direct and sequential search approaches, overcome those shortcomings can be used. The methodology which we have used is based on inferential analysis. To build the models to identify the major future changes in the direction of a situational data series, a comparative model building is applied. Hereby, the paper suggests using some of the nonlinear techniques, like direct and sequential search approaches, to reduce the technical shortcomings. The final result of the paper is to manipulate, to prepare, and to integrate heuristic non-linear searching methods to serve calculating adjusted factors to produce the best forecast data.
文摘Analyzing time series characteristics of red tide is the basis of disaster prevention and mitigation,which is very important to red tide prediction.There are trend comp onents and periodic components in annual time series of occurrence freque ncy and area of red tides,so Gray-Periodic Extensional Combinatorial Model(GPECM)is used to extract these components.The fitting degree of occurrence frequency and area can reach 95.20% and 95.24%,respectively.The performance of GPECM is better than Gray Model,Fourier Series Extension Model,and Holt-Winter Exponential Smoothing Model in model stability.Consequently,it is used to forecast the occurrence frequency and area in 2020 and 2021,and results show that the annual frequency of red tides in 2020 and 2021 can rise to 39 and 41,respectively,and that the annual occurrence area of red tides can rise to 3168 km^(2),which is about 59% more than last year.In 2021,it can fall to 1901 km^(2).
基金Supported by the National Natural Science Foundation of China under Grant No.30600122GuangDong Provincial Natural Science Foundation under Grant No.06025073
文摘Noise can induce inverse period-doubling transition and chaos. The effects of the colored noise on periodic orbits, of the different periodic sequences in the logistic map, are investigated. It is found that the dynamical behaviors of the orbits, induced by an exponentially correlated colored noise, are different in the mergence of transition, and the effects of the noise intensity on their dynamical behaviors are different from the effects of the correlation time of noise. Remarkably, the noise can induce new periodic orbits, namely, two new orbits emerge in the period-four sequence at the bifurcation parameter value μ = 3.5, four new orbits in the period-eight sequence at μ= 3.55, and three new orbits in the period-six sequence at μ = 3.846, respectively. Moreover, the dynamical behaviors of the new orbits clearly show the resonancelike response to the colored noise.
基金supported by the National Natural Science Foundation of China (Grant Nos. 51079046, 50909041, 50809025, 50879024)the National Science and Technology Support Plan (Grant Nos. 2008BAB29B03, 2008BAB29B06)+7 种基金the Special Fund of State Key Laboratory of China (Grant Nos. 2009586012, 2010585212)the Fundamental Research Funds for the Central Universities (Grant Nos. 2009B08514, 2010B20414, 2010B14114)China Hydropower Engineering Consulting Group Co. Science and Technology Support Project (Grant No. CHC-KJ-2007-02)Jiangsu Province "333 High-Level Personnel Training Project" (Grant No. 2017-B08037)the Natural Science Foundation of Hohai University (Grant No. 2008426811)Graduate Innovation Program of Universities in Jiangsu Province (Grant No. CX09B_163Z)the Science Foundation for The Excellent Youth Scholars of Ministry of Education of China (Grant No. 20070294023)Dominant Discipline Construction Program Funded Projects of Universities in Jiangsu Province
文摘A method of the fuzzy cross-correlation factor exponent in dynamics is researched and proposed to diagnose abnormality of cracks in the concrete dam. Moreover, the Logistic time series changing from period-doubling bifurcation to chaos is tested first using this method. Results indicate that it can distinguish inherent dynamics of time series and can detect mutations. Considering that cracks in the concrete dam constitute an open, dissipative and complex nonlinear dynamical system, a typical crack on the downstream face of a concrete gravity arch dam is analyzed with the proposed method. Two distinct mutations are discovered to indicate that the abnormality diagnosis of cracks in the concrete dam is achieved dynamically through this method. Furthermore, because it can be directly utilized in the measured crack opening displacement series to complete abnormality diagnosis, it has a good prospect for practical applications.
