Coupled Attribute Similarity Learning on Categorical Data for Multi-Label Classification

In this paper a novel coupled attribute similarity learning method is proposed with the basis on the multi-label categorical data (CASonMLCD). The CASonMLCD method not only computes the correlations between different attributes and multi-label sets using information gain, which can be regarded as the important degree of each attribute in the attribute learning method, but also further analyzes the intra-coupled and inter-coupled interactions between an attribute value pair for different attributes and multiple labels. The paper compared the CASonMLCD method with the OF distance and Jaccard similarity, which is based on the MLKNN algorithm according to 5 common evaluation criteria. The experiment results demonstrated that the CASonMLCD method can mine the similarity relationship more accurately and comprehensively, it can obtain better performance than compared methods.     

Scheduling Optimization of Space Object Observations for Radar

An optimizing method of observation scheduling based on time-division multiplexing is proposed in this paper,and its efficiency is verified by outdoor experiments.The initial observation scheduling is first obtained by using a semi-random search algorithm,and secondly the connection time pair(CTP) between adjacent objects is optimized by using a genetic algorithm.After obtaining these two parameters,the fimal observation scheduling can be obtained.According to pre-designed tracks between each adjacent objects in observation order,the seamless observation of neighboring targets is derived by automatically steering the antenna beam,so the observation efficiency is improved.     

Scheduling Optimization of Space Object Observations for Radar

An optimizing method of observation scheduling based on time-division multiplexing is proposed in this paper,and its efficiency is verified by outdoor experiments.The initial observation scheduling is first obtained by using a semi-random search algorithm,and secondly the connection time pair(CTP) between adjacent objects is optimized by using a genetic algorithm.After obtaining these two parameters,the final observation scheduling can be obtained.According to pre-designed tracks between each adjacent objects in observation order,the seamless observation of neighboring targets is derived by automatically steering the antenna beam,so the observation efficiency is improved.     

Study and application of time series forecasting based on rough set and Kernel method

A support vector machine time series forecasting model based on rough set data preprocessing was proposed by combining rough set attribute reduction and support vector machine regression algorithm. First, remove the redundant attribute for forecasting from condition attribute by rough set method; then use the minimum condition attribute set obtained after the reduction and the corresponding initial data, reform a new training sample set which only retain the important attributes influencing the forecasting accuracy; study and train the support vector machine with the training sample obtained after reduction, and then input the reformed testing sample set according to the minimum condition attribute and corresponding initial data. The model was tested and the mapping relation was got between the condition attribute and forecasting variable. Eventually, power supply and demand were forecasted in this model. The average absolute error rates of power consumption of the whole society and yearly maximum load are respectively 14.21% and 13.23%. It shows that RS-SVM time series forecasting model has high forecasting accuracy.     

Recommending Personalized POIs from Location Based Social Network

Location based social networks (LBSNs) provide location specific data generated from smart phone into online social networks thus people can share their points of interest (POIs).POI collections are complex and can be influenced by various factors,such as user preferences,social relationships and geographical influence.Therefore,recommending new locations in LBSNs requires to take all these factors into consideration.However,one problem is how to determine optimal weights of influencing factors in an algorithm in which these factors are combined.The user similarity can be obtained from the user check-in data,or from the user friend information,or based on the different geographical influences on each user's check-in activities.In this paper,we propose an algorithm that calculates the user similarity based on check-in records and social relationships,using a proposed weighting function to adjust the weights of these two kinds of similarities based on the geographical distance between users.In addition,a non-parametric density estimation method is applied to predict the unique geographical influence on each user by getting the density probability plot of the distance between every pair of user's check-in locations.Experimental results,using foursquare datasets,have shown that comparisons between the proposed algorithm and the other five baseline recommendation algorithms in LBSNs demonstrate that our proposed algorithm is superior in accuracy and recall,furthermore solving the sparsity problem.     

Recommending Personalized POIs from Location Based Social Network

Location based social networks (LBSNs) provide location specific data generated from smart phone into online social networks thus people can share their points of interest (POIs).POI collections are complex and can be influenced by various factors,such as user preferences,social relationships and geographical influence.Therefore,recommending new locations in LBSNs requires to take all these factors into consideration.However,one problem is how to determine optimal weights of influencing factors in an algorithm in which these factors are combined.The user similarity can be obtained from the user check-in data,or from the user friend information,or based on the different geographical influences on each user's check-in activities.In this paper,we propose an algorithm that calculates the user similarity based on check-in records and social relationships,using a proposed weighting function to adjust the weights of these two kinds of similarities based on the geographical distance between users.In addition,a non-parametric density estimation method is applied to predict the unique geographical influence on each user by getting the density probability plot of the distance between every pair of user's check-in locations.Experimental results,using foursquare datasets,have shown that comparisons between the proposed algorithm and the other five baseline recommendation algorithms in LBSNs demonstrate that our proposed algorithm is superior in accuracy and recall,furthermore solving the sparsity problem.     

Approach of fault diagnosis based on similarity degree matching distance function

A new diagnosis method based on the similarity degree matching distance function is proposed.This method solves the problem that the traditional fault diagnosis methods based on transition system model cannot deal with the"special state"which cannot match the target states completely.For evaluating the relationship between the observation and the target states,this paper first defines a new distance function based on the viewpoint of energy to measure the distance between two attribute values.After that,all the distances of the attributes in the state vector are used to synthesize the distance between two states.For calculating the similarity degree between two states,a trend evaluation method is developed.It analyzes the main direction of the trend of the state transfer according to the distances between the observation and each target state and their historical records.Applying the diagnosis method to a primary power subsystem of a satellite,the simulation result shows that it is effective.     

Localization and mapping in urban area based on 3D point cloud of autonomous vehicles

In order to meet the application requirements of autonomous vehicles, this paper proposes a simultaneous localization and mapping (SLAM) algorithm, which uses a VoxelGrid filter to down sample the point cloud data, with the combination of iterative closest points (ICP) algorithm and Gaussian model for particles updating, the matching between the local map and the global map to quantify particles'' importance weight. The crude estimation by using ICP algorithm can find the high probability area of autonomous vehicles'' poses, which would decrease particle numbers, increase algorithm speed and restrain particles'' impoverishment. The calculation of particles'' importance weight based on matching of attribute between grid maps is simple and practicable. Experiments carried out with the autonomous vehicle platform validate the effectiveness of our approaches.     

Joint optimization algorithm of sensing thresholds and cooperative users in cooperative detection

Cooperative detection is an effective method to improve the spectrum sensing of Cognitive Radio (CR), and its detection performance can be improved through optimization. An optimization algorithm for cooperative detection based on "OR Rule" which can optimize the detection threshold of each user and the number of cooperative users simultaneously is proposed in this paper. The algorithm, which is based on minimizing the error detection probability, adopts partial fusion to improve the detection performance effectively. The simulation results show that the error detection probability of the proposed algorithm is lower than that of the cooperative detection algorithm with the settled threshold, and the better performance can be achieved through choosing fewer users.     

Using frame correlation algorithm in a duration distribution based hidden Markov model

The assumption of frame independence is a widely known weakness of traditional hidden Markov model (HMM). In this paper, a frame correlation algorithm based on the duration distribution based hidden Markov model (DDBHMM) is proposed. In the algorithm, an AR model is used to depict the low pass effect of vocal tract from which stems the inertia leading to frame correlation. In the preliminary experiment of middle vocabulary speaker dependent isolated word recognition, our frame correlation algorithm outperforms the frame independent one. The average error reduction is about 20% .     

基于信息熵的属性约简算法的研究与实现

属性约简是粗糙集理论研究的重要内容之一,是在保持信息系统分类能力不变的基础上,删除冗余属性.为了获得决策系统中属性最小相对约简,本文将信息论应用于决策信息系统属性约简中,与遗传算法相结合,并采用加权平均的属性重要度和知识量作为启发式信息指导约简,提出了一种改进的基于核子集的属性约简算法.     

基于粗集理论的约简算法

在基于属性重要性和基于分辨矩阵两种算法的基础上,提出了一种同时满足属性重要性和频度的启发式约简算法RedFreSigni。该算法的基本思想是:以属性的核为基础,把核和用户偏好集同时作为属性近似约简的一部分,以频度作为选择属性的启发信息可同时生成计算属性的频度信息与不可分辨矩阵,减少了计算时间。在此基础上进而提出了基于规则支持度和置信度的决策挖掘算法,该算法能有效提取出用户感兴趣的规则。     

基于决策规则的属性约简算法

属性约简是粗糙集的核心问题之一。本文基于决策规则给出属性约简相关结论和属性重要性,提出启发式约简算法,引入黄金分割法思想,提高算法效率,并以实例验证算法有效性和正确性。     

信息系统的一种启发式属性约简算法

针对文献[8]中加权平均属性重要度中权值人为确定的不足,提出改进的属性重要度定义,并以实例说明其应用情况.提出约简质量的定义,从属性约简率和近似质量两方面来衡量约简效果.基于改进的属性重要度定义（标准）,构造了两种启发式属性约简算法,并利用UCI数据库中的一些典型算例验证了算法的有效性;说明在某些情况下,提出的属性约简算法在一定程度上能够提高数据的约简质量.     

基于最大外权重的一种启发式属性约简算法

作为属性对区分元素类别所做贡献大小的度量,定义属性重要性权重,并作为启发性知识构造一种寻求约简的启发式算法。算法复杂度是多项式的,收敛速度快,可得到Pawlak约简。     

基于变精度粗糙集模型的属性约简方法研究

提出了一种对存在噪声和不完整数据的决策系统在变精度粗糙集模型下进行属性最小相对约简的方法,将由属性对分类的影响程度和β近似精度共同定义的属性重要性度量作为启发式信息引入遗传算法,通过修正操作算子修复个体,使得个体所对应的属性子集的分类能力不变;修正操作算子中对各属性的属性重要性使用贪心策略进行局部寻优.对遗传算法的各操作算子进行优化,既保证遗传操作过程中种群的多样性,又保证算法能快速收敛.最后通过实例验证了算法的有效性.     

一种改进的粗糙集属性约简启发式算法

作为数据挖掘的重要工具,粗糙集理论被广泛的应用于关系数据库中属性相关性描述、属性集约简、属性重要性度量、规则发现等方面。该文在分析基于信息系统的粗糙集理论的基础上,对基于分辨矩阵的属性约简算法进行了详尽的描述。针对该算法存在的时间和空间性能不理想问题,提出度量单个条件属性对系统概念贡献程度的关联度的概念,以此作为启发式信息对原算法进行改进,得到条件属性的约简。理论分析及实验结果表明该算法具有较好的约简效果及更高的运行效率,为粗糙集理论更广泛地应用于具体的实践提供了一种方法。     

一种基于互信息增益率的新属性约简算法

为了获得决策系统中更好的相对属性约简，提出了一种基于互信息增益率的属性约简算法.该算法考虑了所选择条件属性与决策属性的互信息，还考虑了所选择属性的值的分布情况，从信息论角度定义了基于互信息增益率的属性重要性度量方法，并以此度量为启发式信息，算法从空集开始逐步将最重要的条件属性加入到选择属性集，直到所选择的条件属性集与决策属性集的互信息等于整个条件属性集与决策属性集的互信息时，算法停止.结果表明，算法能更有效地对决策系统进行约简，同时约简后的对象数目较少.     

基于条件信息量的知识相对约简算法

通过在信息系统中引入了知识的条件信息量的概念，证明了在知识相对约简过程中条件信息量的变化趋势是递减的；通过条件信息量定义了属性的相对重要性，提出了一种基于条件信息量的知识相对约简算法，分析得到该算法的时间复杂性为O(|C|^3|U|^2)；通过例子分析，表明该算法是有效的．     

基于粗糙集相关矩阵的属性约简算法

利用粗糙集相关矩阵采用贪婪策略构造了寻找最小属性约简的启发式算法，证明了算法的正确性并作了复杂性分析，通过实例和与基于属性频率重要性算法进行的对比分析，发现该文算法能快速逼近最小约简，且获得的知识容易理解。