Fuzzy Weighted Clustering Method for Numerical Attributes of Communication Big Data Based on Cloud Computing

It is necessary to optimize clustering processing of communication big data numerical attribute feature information in order to improve the ability of numerical attribute mining of communication big data, and thus a big data clustering algorithm based on cloud computing was proposed. The cloud extended distributed feature fitting method was used to process the numerical attribute linear programming of communication big data, and the mutual information feature quantity of communication big data numerical attribute was extracted. Combined with fuzzy C-means clustering and linear regression analysis, the statistical analysis of big data numerical attribute feature information was carried out, and the associated attribute sample set of communication big data numerical attribute cloud grid distribution was constructed. Cloud computing and adaptive quantitative recurrent classifiers were used for data classification, and block template matching and multi-sensor information fusion were combined to search the clustering center automatically to improve the convergence of clustering. The simulation results show that, after the application of this method, the information fusion performance of the clustering process was better, the automatic searching ability of the data clustering center was stronger, the frequency domain equalization control effect was good, the bit error rate was low, the energy consumption was small, and the ability of fuzzy weighted clustering retrieval of numerical attributes of communication big data was effectively improved.


Introduction
With the development of the economy, the progress of science and technology, and the improvement of talent technology, the development of wireless network mobile communication technology has been greatly promoted. With the increasing demand for internet communication, the process of improving the data transmission performance of wireless network mobile communication can help people to have a more comprehensive and clear understanding of the world, to pay attention to the dynamics of the world in real time, and to understand a more abundant amount of knowledge. In the new era, people cannot be separated from wireless network mobile communication data transmission. Whether in work or in daily life, wireless network mobile communication data transmission has penetrated into every corner and has promoted people's interaction. Wireless network mobile communication data transmission depends on optical cable optical fiber. Therefore, the superior performance of optical cable optical fiber communication transmission represents the top economic development and efficiency in the world [1]. In our country, at the end of 2017, the number of optical fiber access (FTTH/0) ports According to Figure 1, the output state characteristic quantity of the communication large data value attribute distribution set in the B model is x j = x 1 j , x 2 j, . . . , x mj T , the sampling is carried out at a baud rate of more than two times, and the state characteristic distribution of the characteristic information of the communication large data value attribute characteristic information is p(x 0 ). The joint feature mining results of the association rules of the text data are as follows: According to the symbol characteristic quantity of communication big data numerical attribute transmission, the information is reconstructed, and the bit sequence distribution of communication big data numerical attribute multi-dimensional big data transmission is obtained by using fuzzy data clustering analysis technology: The semantic concept set of communication big data numerical attribute characteristic information is obtained, and rough set scheduling and frequent mining are carried out for the characteristic information of communication big data numerical attribute. According to the hierarchical characteristics of data aggregation tree, the classification state characteristic quantity of communication big data numerical attribute information is z(t), and the rough concept distribution subset of data clustering center S i (i = 1, 2, · · · , L) that meets the convergence condition of semi-supervisory learning is According to the above analysis, a grid clustering method is used to classify the communication big data numerical attribute feature information, and the small disturbance suppression method is combined to avoid the cluster center disturbance and improve the convergence of the clustering [16][17][18][19][20]. The specific schematic frame diagram is shown in Figure 2: According to the symbol characteristic quantity of communication big data numerical attribute transmission, the information is reconstructed, and the bit sequence distribution of communication big data numerical attribute multi-dimensional big data transmission is obtained by using fuzzy data clustering analysis technology: The semantic concept set of communication big data numerical attribute characteristic information is obtained, and rough set scheduling and frequent mining are carried out for the characteristic information of communication big data numerical attribute. According to the hierarchical characteristics of data aggregation tree, the classification state characteristic quantity of communication big data numerical attribute information is () zt , and the rough concept distribution subset of data clustering center ( 1, 2, , ) i S i L  that meets the convergence condition of semisupervisory learning is According to the above analysis, a grid clustering method is used to classify the communication big data numerical attribute feature information, and the small disturbance suppression method is combined to avoid the cluster center disturbance and improve the convergence of the clustering [16][17][18][19][20]. The specific schematic frame diagram is shown in Figure 2:

Communication Big Data Numerical Attribute Linear Programming Processing
The cloud extended distributed feature fitting method is used to process the numerical attribute linear programming of communication big data, and the mutual information feature quantity of big data numerical attribute of communication is extracted, which is described as follows:

Communication Big Data Numerical Attribute Linear Programming Processing
The cloud extended distributed feature fitting method is used to process the numerical attribute linear programming of communication big data, and the mutual information feature quantity of big data numerical attribute of communication is extracted, which is described as follows: The scalar time series of communication big data numerical attribute feature information is x(t), t = 0, 1, · · · , n − 1, given the information flow of communication big data numerical attribute feature information. Given the vector group x 1 , x 2 , · · · x n ∈ C m (m-dimensional complex space), combined with the linear programming method, the finite set of communication big data numerical attribute feature information set distribution is obtained as follows: Σ = diag max |ρ + 1 |, |ρ − 1 | , · · · max |ρ + n |, |ρ − n | = diag ρ 1 , · · · ρ n (5) The fault location process of communication big data numerical attribute fuzzy weighted clustering is shown in Figure 3: The scalar time series of communication big data numerical attribute feature information is () xt , 0,1, , 1 tn    , given the information flow of communication big data numerical attribute feature information. Given the vector group 12 The fault location process of communication big data numerical attribute fuzzy weighted clustering is shown in Figure 3: The intelligent distributed feeder system automation management is the basic work for realizing the whole communication big data system automation management and power fault early warning, identification, and elimination. When the feeder system breaks down, it can quickly locate and isolate the fault points by using the system distributed topology structure and equation communication mode to restore the normal power supply in the non-fault area of the power grid quickly. The overall analysis of the distributed feeder topology system structure based on the peer-to-peer communication mainly includes the base station layer, the sub station layer and the terminal layer. The peer-to-peer network connection is used between each layer, as shown in Figure 4. The intelligent distributed feeder system automation management is the basic work for realizing the whole communication big data system automation management and power fault early warning, identification, and elimination. When the feeder system breaks down, it can quickly locate and isolate the fault points by using the system distributed topology structure and equation communication mode to restore the normal power supply in the non-fault area of the power grid quickly. The overall analysis of the distributed feeder topology system structure based on the peer-to-peer communication mainly includes the base station layer, the sub station layer and the terminal layer. The peer-to-peer network connection is used between each layer, as shown in Figure 4. According to the different functions of nodes, the topology of the feeder system is divided into base station layer, sub station layer, and terminal layer. However, each node has the same position in the feeder system, which is not a subordinate relationship but a distributed parallel relationship. The nodes in the terminal layer are subdivided into feeder monitoring terminal (FTU), switching terminal (DTU), and distribution terminal (TTU). The base station layer acts as the control center of the system from the structural design, and is mainly responsible for the system broadcast and information summary of the power fault information. Similarly, the sub station layer nodes and the terminal layer nodes also have the system broadcast function of the fault information in the functional design; the nodes in the terminal layer are mainly responsible for the identification, measurement, and control of the fault within the area, and the terminal node itself has the function of the system broadcast. It can also package and transfer the node information to the sub station layer to improve the efficiency of power grid information transmission.
In the equal communication mode, the feeder topology system adopts the modular design, mainly including the central processing module, power module, bus communication module, switch module, and display module. The main module adopts the stm32h7 single chip and 32-bit bus system, and matches the hl2323ds communication chip. This structure design can not only meet the requirements of the feeder system topology on communication mode, but also improve the power fault information throughput of the distributed feeder system. The hardware connection design of the stm32h7 single chip microcomputer is shown in Figure 5.   The stm32h7 single chip microcomputer is responsible for the control and information processing functions of the distributed feeder system. The chip is connected with other system modules through RS485 bus and can bus, and receives and transmits fault information of the feeder system, analog system calculation, control switch, and analysis message, among others. The stm32h7 single chip microcomputer has good compatibility and rich interfaces. RS232 communication interface is selected as the bus communication module as the remote data exchange window with the According to the different functions of nodes, the topology of the feeder system is divided into base station layer, sub station layer, and terminal layer. However, each node has the same position in the feeder system, which is not a subordinate relationship but a distributed parallel relationship. The nodes in the terminal layer are subdivided into feeder monitoring terminal (FTU), switching terminal (DTU), and distribution terminal (TTU). The base station layer acts as the control center of the system from the structural design, and is mainly responsible for the system broadcast and information summary of the power fault information. Similarly, the sub station layer nodes and the terminal layer nodes also have the system broadcast function of the fault information in the functional design; the nodes in the terminal layer are mainly responsible for the identification, measurement, and control of the fault within the area, and the terminal node itself has the function of the system broadcast. It can also package and transfer the node information to the sub station layer to improve the efficiency of power grid information transmission.
In the equal communication mode, the feeder topology system adopts the modular design, mainly including the central processing module, power module, bus communication module, switch module, and display module. The main module adopts the stm32h7 single chip and 32-bit bus system, and matches the hl2323ds communication chip. This structure design can not only meet the requirements of the feeder system topology on communication mode, but also improve the power fault information throughput of the distributed feeder system. The hardware connection design of the stm32h7 single chip microcomputer is shown in Figure 5. According to the different functions of nodes, the topology of the feeder system is divided into base station layer, sub station layer, and terminal layer. However, each node has the same position in the feeder system, which is not a subordinate relationship but a distributed parallel relationship. The nodes in the terminal layer are subdivided into feeder monitoring terminal (FTU), switching terminal (DTU), and distribution terminal (TTU). The base station layer acts as the control center of the system from the structural design, and is mainly responsible for the system broadcast and information summary of the power fault information. Similarly, the sub station layer nodes and the terminal layer nodes also have the system broadcast function of the fault information in the functional design; the nodes in the terminal layer are mainly responsible for the identification, measurement, and control of the fault within the area, and the terminal node itself has the function of the system broadcast. It can also package and transfer the node information to the sub station layer to improve the efficiency of power grid information transmission.
In the equal communication mode, the feeder topology system adopts the modular design, mainly including the central processing module, power module, bus communication module, switch module, and display module. The main module adopts the stm32h7 single chip and 32-bit bus system, and matches the hl2323ds communication chip. This structure design can not only meet the requirements of the feeder system topology on communication mode, but also improve the power fault information throughput of the distributed feeder system. The hardware connection design of the stm32h7 single chip microcomputer is shown in Figure 5.  The stm32h7 single chip microcomputer is responsible for the control and information processing functions of the distributed feeder system. The chip is connected with other system modules through RS485 bus and can bus, and receives and transmits fault information of the feeder system, analog system calculation, control switch, and analysis message, among others. The stm32h7 single chip microcomputer has good compatibility and rich interfaces. RS232 communication interface is selected as the bus communication module as the remote data exchange window with the The stm32h7 single chip microcomputer is responsible for the control and information processing functions of the distributed feeder system. The chip is connected with other system modules through RS485 bus and can bus, and receives and transmits fault information of the feeder system, analog system calculation, control switch, and analysis message, among others. The stm32h7 single chip microcomputer has good compatibility and rich interfaces. RS232 communication interface is selected as the bus communication module as the remote data exchange window with the sub station layer and the terminal layer. The operation information of each module of the distributed feeder system is displayed more intuitively through the display module. The stable operation of the distributed feeder system based on the equal communication requires a reliable power supply module. Because the design of the feeder system is mainly aimed at the overhead line, the AC power supply is selected to facilitate the stable operation of the information control system. Under the normal working condition, the working power supply of the distributed feeder topology structure is also responsible for the stable power supply of the PT line. The power module is also equipped with a battery as the standby power supply. When the AC power supply fails, the battery can guarantee the normal operation of the system for a short time.
The piecewise sample combination design of the fusion data is carried out [21,22], and when the time interval of association rule set feature extraction of communication big data numerical attribute feature information is O(d) of O(N 1 /d ), data clustering space snξ → tanhξ obtains validation that the boundary value convergence condition of accurate clustering of communication big data numerical attribute feature information is satisfied.
Three kinds of kernel functions have been designed to represent the linear kernel function, random distribution characteristic kernel function, and uniform distribution kernel function of big data numerical attribute feature information clustering. The expressions are, respectively, as follows: According to the above three kernel functions, the linear programming design of accurate clustering of big data numerical attribute feature information is carried out, and the convergence control ability in the process of data clustering is improved by combining the semi-supervised learning algorithm.

Big Data Fuzzy Weighted Clustering Optimization
On the basis of using cloud extended distributed feature fitting method to deal with big data numerical attribute linear programming, the optimal design of big data fuzzy weighted clustering algorithm is carried out. In this paper, a big data fuzzy weighted clustering algorithm based on cloud computing was proposed. The mutual information feature quantity of communication big data numerical attribute is extracted, and the characteristic distribution value of geometric neighborhood (t, f ) of communication big data numerical attribute clustering in nonlinear space is obtained.
Combined with fuzzy C-means clustering and linear regression analysis, the statistical analysis of numerical attribute feature information of communication big data is carried out. In the clustering space matrix (x 1 , x 2 , · · · , x n ), the basis vector G = [E k×k |A] of data fuzzy weighted clustering is obtained in order to construct the joint disturbance feature equation group of communication big data numerical attribute feature information clustering.
On the basis of the above analysis of the boundary value convergence conditions for accurate clustering of numerical attribute characteristic information of communication big data, the stable convergence of the fuzzy weighted clustering mathematical model of the whole data is guaranteed. Using a semi-supervised learning method, the boundary solution vector function of numerical attribute feature information clustering of communication big data is constructed as follows: Combined with fuzzy C-means clustering and linear regression analysis [23,24], the statistical analysis of big data's numerical attribute feature information is carried out, and the statistical feature equation is described as follows: . (16) in which . In order to realize data optimization clustering, a new training vector is input in a finite-dimensional space: For the method, the convergence constraint control of the iterative process is carried out by adopting the cloud computing and the adaptive quantitative recursive analysis, and the spatial clustering of the data fuzzy weighted clustering center is obtained as follows: wherein, ω j = (ω 0 j , ω 1 j, · · · , ω k−1, j ) T . The fault location algorithm based on the network topology is used to form the topology tree structure of all nodes in the intelligent feeder network, and the cluster center is selected according to the optimal location of the nodes to form a number of node clusters [25][26][27][28]. During fault location, firstly traverse the cluster center node of each cluster. When the information is obtained from the cluster center node, there is abnormal node state in the cluster; then, traverse all nodes in the cluster, and then identify the location of the fault node and isolate the fault area, and synchronously transfer the fault information to other cluster center nodes [29][30][31][32]. When the clustering center satisfies the convergence condition of the semi-supervised learning, the detection statistic of the characteristic information of the large-data value of the communication satisfies the clustering convergence condition, and the implementation process of the large-data fuzzy weighted clustering algorithm designed in this paper is obtained, as shown in Figure 6.  Figure 6. Data clustering output.

Simulation Experiment Analysis
In order to test the performance of this method in the clustering of numerical attribute information of communication big data and verify the effectiveness and feasibility of this method, simulation experiments were carried out. On the basis of the Deep Web database and MATLAB, we empirically designed a fuzzy weighted clustering algorithm. The attribute of the big data sample was set to 6, and the initial confidence level of data fuzziness was. The clustering was 95%, the critical value was 1.24, and the judgment threshold was 0.13. The embedding dimension of the feature space distribution was set to M = 4, the data length of the test sample was set at2000 years, and the simulation time was120s. The method of this paper was used to compare study [2] and study [3] for experimental and comparative analysis. The main parameters of the experiment are shown in Table  1. According to the above simulation environment and parameter settings, big data fuzzy weighted clustering analysis communication big data numerical attribute Mu executed onedimensional text, and the original data distribution is shown in Figure 7.

Simulation Experiment Analysis
In order to test the performance of this method in the clustering of numerical attribute information of communication big data and verify the effectiveness and feasibility of this method, simulation experiments were carried out. On the basis of the Deep Web database and MATLAB, we empirically designed a fuzzy weighted clustering algorithm. The attribute of the big data sample was set to 6, and the initial confidence level of data fuzziness was. The clustering was 95%, the critical value was 1.24, and the judgment threshold was 0.13. The embedding dimension of the feature space distribution was set to M = 4, the data length of the test sample was set at2000 years, and the simulation time was120s. The method of this paper was used to compare study [2] and study [3] for experimental and comparative analysis. The main parameters of the experiment are shown in Table 1. According to the above simulation environment and parameter settings, big data fuzzy weighted clustering analysis communication big data numerical attribute Mu executed one-dimensional text, and the original data distribution is shown in Figure 7. Symmetry 2018, 10, x FOR PEER REVIEW 10 of 14   Taking the data of Figure 7 as the research object, the fuzzy weighted clustering of the data was carried out, with the data classification being carried out by using cloud computing and adaptive quantitative recurrent classifiers. The clustering output results are shown in Figure 8.  Figure 8 shows that big data fuzzy weighted clustering can be effectively realized by using this method; the accuracy of data classification was high, and the error rate was small. The performance of big data fuzzy weighted clustering was tested by different methods.
The comparison results are shown in Figure 9. It can be seen from Figure 9 that the fuzzy weighted clustering error rate of big data was lower than that of the other two comparison methods after adopting this method, which proved that this method has obvious application advantages. Taking the data of Figure 7 as the research object, the fuzzy weighted clustering of the data was carried out, with the data classification being carried out by using cloud computing and adaptive quantitative recurrent classifiers. The clustering output results are shown in Figure 8. Taking the data of Figure 7 as the research object, the fuzzy weighted clustering of the data was carried out, with the data classification being carried out by using cloud computing and adaptive quantitative recurrent classifiers. The clustering output results are shown in Figure 8.  Figure 8 shows that big data fuzzy weighted clustering can be effectively realized by using this method; the accuracy of data classification was high, and the error rate was small. The performance of big data fuzzy weighted clustering was tested by different methods.
The comparison results are shown in Figure 9. It can be seen from Figure 9 that the fuzzy weighted clustering error rate of big data was lower than that of the other two comparison methods after adopting this method, which proved that this method has obvious application advantages.  Figure 8 shows that big data fuzzy weighted clustering can be effectively realized by using this method; the accuracy of data classification was high, and the error rate was small. The performance of big data fuzzy weighted clustering was tested by different methods.
The comparison results are shown in Figure 9. It can be seen from Figure 9 that the fuzzy weighted clustering error rate of big data was lower than that of the other two comparison methods after adopting this method, which proved that this method has obvious application advantages. In order to test the effect of frequency domain equalization of communication security, this paper compared study [2] and study [3], and compared the performance of this method in the routing protocol that uses node location information to make decisions, finding that the performance of this method was better in the network with high dynamic topology caused by node movement. The In order to test the effect of frequency domain equalization of communication security, this paper compared study [2] and study [3], and compared the performance of this method in the routing protocol that uses node location information to make decisions, finding that the performance of this method was better in the network with high dynamic topology caused by node movement. The comparison results of the average energy consumption of the three methods for successful transmission of a single packet are shown in Figure 10. In order to test the effect of frequency domain equalization of communication security, this paper compared study [2] and study [3], and compared the performance of this method in the routing protocol that uses node location information to make decisions, finding that the performance of this method was better in the network with high dynamic topology caused by node movement. The comparison results of the average energy consumption of the three methods for successful transmission of a single packet are shown in Figure 10. It can be seen from Figure 10 that in the comparison of the three methods for successfully receiving the average energy consumption of a single packet, the unit energy consumption of study [2] and study [3] was more, and the text energy consumption was the least, and thus the method in this paper is the best for the communication security frequency-domain equalization control effect. When the communication channel was weak, its noise had a great influence on the capacity of the communication channel. However, when the minimum mean square error equalization was used, the optimal energy constraint and frequency-domain equalization were controlled to ensure the optimal energy in the frequency domain equalization control, and the average energy consumption control of a single data packet was successfully transmitted. The communication channel information obtained by channel estimation was equalized in the frequency domain. The communication channel had three Rayleigh paths, and the specific communication channel equalization energy control results are shown in Figure 11. It can be seen from Figure 10 that in the comparison of the three methods for successfully receiving the average energy consumption of a single packet, the unit energy consumption of study [2] and study [3] was more, and the text energy consumption was the least, and thus the method in this paper is the best for the communication security frequency-domain equalization control effect. When the communication channel was weak, its noise had a great influence on the capacity of the communication channel. However, when the minimum mean square error equalization was used, the optimal energy constraint and frequency-domain equalization were controlled to ensure the optimal energy in the frequency domain equalization control, and the average energy consumption control of a single data packet was successfully transmitted. The communication channel information obtained by channel estimation was equalized in the frequency domain. The communication channel had three Rayleigh paths, and the specific communication channel equalization energy control results are shown in Figure 11. It can be seen from Figure 11 that the bit error rate of the communication channel control conducted by study [2] and study [3] was higher, whereas that of the communication channel equalization energy control conducted by this method was lower, showing a downward trend. Using It can be seen from Figure 11 that the bit error rate of the communication channel control conducted by study [2] and study [3] was higher, whereas that of the communication channel equalization energy control conducted by this method was lower, showing a downward trend. Using this method to balance the energy of the communication channel can optimize the system well, and its security performance is good. Because of the small energy loss, the life of communication network can be greatly extended.
From what has been discussed above, the information fusion performance of big data numerical attribute information clustering processing was better, the automatic search ability of data clustering center was stronger, and the fuzzy weighted clustering retrieval ability of communication big data numerical attribute was improved. This method had good application value in cloud computing analysis and clustering of communication data.

Conclusions
In this paper, a big data clustering algorithm based on cloud computing was proposed. The cloud extended distributed feature fitting method was used to process the numerical attribute linear programming of communication big data, and the mutual information feature quantity of communication big data numerical attribute was extracted. Combined with fuzzy C-means clustering and linear regression analysis, the statistical analysis of big data numerical attribute feature information was carried out, and the associated attribute sample set of communication big data numerical attribute cloud grid distribution was constructed. Cloud computing and adaptive quantitative recurrent classifiers were used for data classification, and block template matching and multi-sensor information fusion were combined to search the clustering center automatically to improve the convergence of clustering. The simulation results showed that the information fusion performance of big data numerical attribute feature information clustering processing was better, and the automatic search ability of data clustering center was strong, which improved the fuzzy weighted clustering retrieval ability of communication big data numerical attributes. This method had good application value in cloud computing analysis and clustering of communication data.
Author Contributions: H.D. and J.Z. a big data clustering algorithm based on cloud computing is proposed in this study. H.D. processed the numerical attribute linear programming of communication big data, and the mutual information feature quantity of communication big data numerical attribute was extracted by the cloud extended distributed feature fitting method. J.Z. and C.S. Combined with fuzzy C-means clustering and linear regression analysis, the statistical analysis of big data numerical attribute feature information was carried out, and the associated attribute sample set of communication big data numerical attribute cloud grid distribution was constructed. H.D. and J.Z. did the experiments, recorded data, and created manuscripts. All authors read and approved the final manuscript.