Design and analysis of management platform based on financial big data

Problem description

The grouping sample test analysis method is used to analyze the correlation of financial data, and the test set is associated with the category characteristics of business process information. The space–time norm of automatic classification of financial data satisfies the following boundary conditions: (1)${\displaystyle \left\{\begin{array}{c}f:R^n\times R^n\times S\hfill \\ g:R^n\times R^n\times S\to R^{n\times d}\hfill \\ h:R^n\times R^n\times S\hfill \end{array}\right.}$

Taking the fuzzy cluster distribution of business process as the central vector, the multivariate discriminant model is used to predict the number of financial data categories, and the statistical function of financial data classification is obtained as follows: (2)${\displaystyle \begin{array}{cc}\hfill & \hfill h:R^n\times R^n\times S\\ \hfill & \hfill \delta \left(t\right):\left[0,T\right]\to R\\ \hfill & \hfill v\left(dt,du\right)=v\left(dt,du\right)-\pi \left(du\right)dt\end{array}}$

According to the correlation analysis results of financial data, the quality of financial data classification is evaluated, and the combination problem of financial data classification is described as follows:

(1) C([a, b], R) is the classification continuous function of financial data of [a, b] in classification space R, and the cross-term samples of classification attributes meet the following requirements: (3)${\displaystyle -\tau <\delta \left(t\right)<t,|\delta \left(t\right)-\delta \left(s\right)|<p|t-s|}$

(2) In the initial clustering space, the information fusion center C([a, b], R) of financial data classification has upper and lower boundaries, and (Eq. 4) should be satisfied (4)${\displaystyle E|\zeta \left(t\right)-\zeta \left(s\right){|}^2<K{\left(t-s\right)}^2,-\tau <s<t<0.}$

Data classification

Combined with Bayesian judgment method, automatic classification and classification robustness analysis of financial data are realized. Classification template of financial data under business process satisfies positive multi-solutions. f: → R, internal control index α > 0, the linear discriminant function of classification statistics is: (5)${\displaystyle D_0^{\alpha }f\left(t\right)=\frac{1}{\Gamma \left(n-\alpha \right)}{\left(\frac{d}{\mathrm{d}t}\right)}^n{\int }_0^t\frac{1}{{\left(t-s\right)}^{\alpha -n+1}}f\left(s\right)\mathrm{d}s.}$

By using validity test and model fitness test, the characteristic functional of data automatic classification is obtained: (6)${\displaystyle \frac{1}{m}\sum _{i=0}^p{a}_i{m}^i\sum _{k=-q/2}^{q/2}{b}_k{c}_k^i=a_1.}$

Under fuzzy constraints c_k =  − c_−k, if q = 4, b₂ = b₋₂ = 1, b₁ =b₋₁ = 2, b₀ = 0, Cauchy-Had-Amard differential equation is used for statistical analysis of financial business process data, and (Eq. 7) can be obtained. (7)${\displaystyle u\left(t\right)=c_1{t}^{\alpha -1}+c_2{t}^{\alpha -2}+,\dots ,+c_N{t}^{\alpha -N}.}$

The distributed adjacent characteristic quantities of financial business process data are calculated. In the statistical data area, the classification closure function on IR^d is denoted as: (8)${\displaystyle \hat{f}\left(\xi \right)={\left(2\pi \right)}^{-d/2}{\int }_{I{R}^d}{e}^{-ix\cdot \xi }f\left(x\right)\mathrm{d}x}$

if s⩾0, its stationary period satisfies (Eq. 9): (9)${\displaystyle \begin{array}{cc}\hfill {\dot{H}}_x^s\left(I{R}^d\right)& \hfill =\left\{f:\parallel f{\parallel }_{{\dot{H}}_x^{{}^{\prime }}\left(I{R}^{{}^{\prime }}\right)}:={∥|\nabla {|}^{s}f∥}_{L_x^2\left(I{R}^d\right)}\right.\\ \hfill & \hfill \left.={∥|\xi {|}^s\hat{f}∥}_{L_2^2\left(I{R}^d\right)}<\infty \right\}.\end{array}}$

After iterative selection, taking the business process as the identification parameter, the final discriminant function of automatic classification of financial data is obtained as follows: (10)${\displaystyle \parallel f{\parallel }_{L_i^q{Z}_1^{{}^{\prime }}\left(I\times I{R}^k\right)}={\left({\int }_I{\left({\int }_{IR}|f\left(t,x\right){|}^r\mathrm{d}x\right)}^{q/r}\mathrm{d}t\right)}^{1/q}.}$

By observing the initial sample, the classification model is judged, and the automatic classification of financial data is realized by the subsection detection method, and the improved design of classification algorithm is realized.