Sample Density Clustering Method Considering Unbalanced Data Distribution

The data distribution of the multidimensional array sensor is unbalanced in sample collection. To improve clustering ability samples, a density method sparse scattered points and multisensor samples based on analysis characteristics proposed. network’s collection structure created using Voronoi polygon topology. By analyzing parameters between classes reconstructing characteristic space sequence, time series collected by reorganized, statistical quantity high-order cumulant sparsely are extracted. Combined with learning algorithm feature fusion, fuzzy information flow elements realized. According to convergence analysis, detection adopted realize optimization configuration elements. test results show that this paper has good convergence, strong spectrum expansion ability, low error rate when collecting arrays.