Research on the Data Acquisition Method Improvement for Mobile Communication Network Based on Cloud Computing

2018 International Conference on Modeling, Simulation and Optimization (MSO 2018) ISBN: 978-1-60595-542-1

Research on the Data Acquisition Method Improvement for Mobile

Communication Network Based on Cloud Computing

Liu-yang WANG

_{, Yang-xin YU and Lei ZHOU}

Faculty of Computer & Software Engineering, Huaiyin Institute of Technology, China *Corresponding author

Keywords: Cloud computing, Mobile communication network, Data acquisition.

Abstract. Aiming at the problem of low efficiency and poor accuracy of data acquisition when using the traditional acquisition method in mobile communication network in the case of external interference. This paper proposes a data acquisition method for mobile communication network based on cloud computing, the data eigenvector is extracted and effectively identify from the time-frequency distribution collected by the mobile communication network equipment. The ADASYN algorithm is used to remove redundancy information based on cloud computing to accurately capture the data of mobile communication network. The experimental results show that the proposed method can effectively achieve the acquisition of mobile communication network data with high accuracy and efficiency.

Introduction

With the development of mobile communication technology and the national layout of the communications industry, the scale of China's mobile communication network construction has reached the top level in the world [1-3]. At present, the proportion of users in the mobile communication network in China increases rapidly proportionately, which increases the amount of data in the mobile communication network [4-6]. In this case, with the rapid increase of complaint volume of mobile communication networks, the gradual increase in the size of mobile networks and gradual changes in the business of network users, the diversity of mobile network communication data has emerged and the data of the mobile communication network needs to be collected. How to collect the data of mobile communication network effectively and rapidly has become an urgent problem to be solved in this field, attracting the attention of the majority of scholars and many good methods have also emerged.

Literature [9] proposed a method of data acquisition based on linear regression. Constructing communication data model by using linear regression analysis method, maintaining the characteristics of communication data, enabling nodes to transmit only the parameter information of regression model to achieve the collection of communication data, however, there is a problem of incomplete data acquisition. Literature [10] proposed a data acquisition method based on interactive data migration technology, which can clean redundant data and increase the accuracy of data. However, there is a problem of slow data acquisition. Literature [11] proposes a data acquisition method based on thematic crawler, which uses the concept and method of thematic crawler and uses regular expressions to match, effectively increasing the performance of data acquisition, however, there is a problem of inefficient data acquisition.

Research on Data Acquisition Methods in Mobile Communication Network

Mobile Communication Network Data Feature Extraction

Extracting the data features of the mobile communication network can eliminate external interference and more accurately and effectively collect the data of the mobile communication network. Cloud computing has no strict restriction on the user's location. The user can use the terminal system to obtain the required services. Even if the communication distance is far, the cloud network can be used to collect and process the communication network data. Data resources are stored in the cloud instead of the traditional fixed tangible entities. Instead, the data resources are run somewhere under the cloud. In fact, the terminal does not need to know the specific location where the application runs. The data can be processed only by the network, including even supercomputing the task is also available in the cloud to provide powerful computing capabilities at any time, and the use of fault-tolerant mechanisms to maintain the stability of data identification and processing.

The data eigenvector of mobile communication network can be effectively identified based on the time-frequency analysis theory. Time-frequency transform is widely used in the field of data processing, has a wide range of effects on all aspects of data information theory, and made a lot of research results. And it is widely used in the field of singularity detection, image processing and decomposition, computer vision and coding, signal estimation, speech analysis and synthesis, fault diagnosis and other fields in the communication network. It provides a new tool for data acquisition; time-frequency distribution theory has become an important basis for data acquisition and signal processing. Especially based on the time-frequency distribution theory, due to its own characteristics, has become a research hotspot. Some applications and research situations of time-frequency transform theory in data information acquisition and signal processing, in which Cohen time-frequency distribution forms the basis and core of time-time-frequency analysis theory. Smoothing WVD is the most frequently used time-frequency representation in time-frequency analysis. Smoothing WVD has good time-frequency clustering and is a typical Cohen time-frequency distribution. The WVD time-frequency distribution method is used to extract the eigenvectors of communication data can effectively identify and collect the data of the communication network. Firstly, the signal WVD is extracted and the edge frequency distribution of the signal data is calculated and solved. The edge frequency distribution of the mobile communication network data is smoothed and finally the normalized processing is performed. Secondly, the band of the data is calculated. According to the band information, the part containing the data features is truncated from the original time-frequency. Finally, the size of the separated data is adjusted, and the adjusted two-dimensional matrix is transformed into a one-dimensional vector, that is, the eigenvector is sought. By solving the eigenvector, the data acquisition of the mobile communication network is realized.

Extracting data eigenvectors of mobile communication network, the local mean and local variance are estimated by the following formula (1) and formula (2), and analyze the frequency edge vector

b

A（ ） . The signal-to-noise ratio of the mobile communication network is relatively low, and the data

frequency band of the collected data fluctuates greatly. First, the edge frequency distribution of the data in the mobile communication network needs to be smoothed and adaptive filter is used to reduce the noise from the outside. Then, use the average sliding filter to suppress the maximum point and edge effects. In this paper, the Wiener filter is used to adjust the parameters of the mobile communication network according to the statistical characteristics of the input data signal, and the local mean (b) and local variance _2_{(b) are estimated according to the edge frequency} distribution of 2a + 1 elements.

1

(b)= ( )

2a 1

k b a

k b a

A k

  

 





2_(b)= 1 2_{( )} 2_{( )} 2a 1

k b a

k b a

A k k

   

 



New eigenvectors are extracted and transformed based on the calculated mean and variance:

2 2

2 (b)

(b)= b) ( ( ) ( )) (b)

P

A   A b  b

 

 ( 

(3) (b)

A is the frequency edge distribution vector, P2 is the average local variance, and the local

variance presents the opposite trend. When the local variance becomes smaller, the filter performs a more substantial smoothing; when the local variance becomes larger, the filter smoothes the local amplitude. The elimination of interference factors such as ambient noise around the mobile communication network through the filter can improve the clarity of the data and the effectiveness of data acquisition. Will filter the processed output smoothing again, to eliminate local interference, optimize the mobile communication network data acquisition results, and reduce the local pulse interference: 1 = ( ) 1 a avg k b

A A b k

b a 

 

 





(4) Finally, the output of the normalized:

a a max( ) vg vg A A A    (5) Through the formula (4) and formula (5) to reduce the pulse interference of local eigenvectors, the filter is normalized processing done, the collected data to do the threshold processing, the extraction of eigenvectors to achieve mobile communication network data eigenvectors extraction.

Improvement of Data Acquisition Method in Mobile Communication Network

On the basis of extracting the eigenvectors of mobile communication network data, the cloud computing method is introduced to collect the data of the mobile communication network to improve the accuracy of the data of the mobile communication network and reduce the missed detection rate.

For the traditional acquisition methods, there is always the problem of low acquisition accuracy and poor efficiency. A data acquisition method based on cloud computing for mobile communication networks is proposed, which can optimize the collected data. First, the data processing model under steady state should be set up to obtain the normal sample data feature vector. Not only can reflect the characteristics of the conventional system or user behavior model, the characteristic of the abnormal behavior data is compared with the difference of the characteristic in the regular behavior to distinguish the data characteristics of the mobile communication network, the processing method is as follows:

It is assumed that the mobile communication network data feature sample points are represented by xi, and the data feature sample points in the data processing range are taken as a reference, and

are obtained according to the reconstruction error function formula, and the full spacing is defined by the fuzzy rule algorithm to the mobile communication network effective distance , The formula is:

( ) ( )

i j

ij

x x

d

a i a j

 

(6) In the formula, ( )a i and ( )a j represent the average value of the adjacent distance, and then solve

the weight wij between the nearest neighbor points xi in the mobile communication network data,

2

i 1 j=1

( ) n - n _ijy_j ( )

i

Y Y w tr y ay





 

 

(7) Through the formula (7), the weight of each adjacent data point can be solved, the loss of function can be reduced, and the effectiveness of communication data acquisition can be improved. In the process of data acquisition, the redundant information interferes with the identification and processing of normal data. In view of how to deal with the redundant information, the method proposed in this paper considers the reliability, operability, and the actual situation, using ADASYN algorithm based on the redundant data itself The distribution of features, generate redundant data sets, and delete then.

First, calculate the number of data samples to be generated for the entire subclass sample using the following formula:





 

 N N

S _{maj (8)}

Second, for each small sample, give its K-nearest neighbor according to the Euclidean distance and compute the density of the redundant data:

min ,..., 2 , 1

,i N

KZ A_i

i  



(9) In the formula, A_i is the sample number of the K nearest neighbors of class i sample and Zis a

constant, then it can be determined that the amount of redundant data is:

S

n_i_i ₍₁₀₎

If the following conditions are met, the redundant data can be deleted to increase the accuracy of data acquisition. The constraints are as follows:

       1 0 0 1 S S

 ₍₁₁₎ Only when S 1_{, for the redundant data removal results of the optimal state, on this basis, the}

cloud computing method for mobile communication network data acquisition results can be expressed as follows:

2

i 1 j=1

- x ( )

n n

ij ij j ij

i

w X w d Y





 

(12) The algorithm uses the time-frequency analysis theory to identify and extract the data eigenvectors of the mobile communication network, excludes external interference, and completes the collection of mobile communication network data based on cloud computing.

Test Results and Analysis

Figure 1. The data frequency distribution of the literature [3] method.

Figure 2. The data frequency distribution of the method in this paper.

Compared to Fig.1 and Fig.2, data acquisition by this paper method can significantly reduce external interference. Fig. 1 shows that the method in literature [3] in collecting the data of the mobile communication network is obvious, which will lead to the reduction of the intensity and clarity of the data signal. The proposed method in this paper can effectively reduce the impact of interference from outsiders, as shown in Fig.2.

The data acquisition method based on cloud computing in mobile communication network proposed in this paper can still maintain high recognition probability of communication data under the condition of low signal-to-noise ratio (SNR).

Figure 3. Recognition probability.

It can be seen from Fig.3 that the proposed method in this paper can still obtain a higher correct recognition probability when the SNR is 4dB.

The proposed algorithm in this paper can find clusters of arbitrary shape in noisy data sets, and effectively solve the algorithm problems in literature [4]. The high-density clustering results of literature [4] algorithm are completely contained in the connected low-density clustering results. Compared with different datasets as shown in Fig.4 and Fig.5, the proposed algorithm in this paper can faithfully reflect the distribution of data in the dataset and can reduce the interference factor.

In this paper, the acquisition of the mobile communication network data by computing method based on the cloud, to find out the data communication network minimum support threshold and minimum confidence threshold conditions of fuzzy association rules in the process of mining, and through a large number of complex and accurate cloud calculation, complete the acquisition of mobile communication network data. The superiority of proposed algorithm in this paper can be obtained by comparing 500 groups sample data with the algorithm in literature [4].

Table 1. Comparison and Analysis of Experimental Data.

Test Method Leakage rate The method in literature [5] 7%

The method in this paper 2%

Figure 6. Data Recognition Accuracy.

From the test data of Table 1, we can see that the accuracy of mobile communication network data detection is only 75% in literature [5], and the accuracy of the method adopted in this paper can reach 92%.

Table 2. Comparison and Analysis of Experimental Data.

Test Method Accuracy Rate The method in literature [5] 75%

The method in this paper 92%

Figure 7. The Data recognition leakage rate.

It is impossible to get all the ideal data in the actual data acquisition and processing. The data we get often contain some interference data, and the processing of communication data can’t reach the ideal state, so the algorithm has some robustness is very important. Based on the experimental data, he robustness curves of data processing under the traditional method and the algorithm in this paper are drawn, as shown in Fig. 8 and in Fig. 9.

Figure 8. The Robustness of Literature [4] Method.

Figure 9. The Robustness of Neural Network.

As shown in Fig.8 and Fig.9, the dotted line is a robust curve for the ideal state data processing. Under the traditional method, the actual robust curve is obviously deviate from the curve under the ideal state. In this algorithm, the actual robustness curve and the ideal state curve still have a deviation, and the robustness is greatly improved when compared with the traditional method.

Experiments show that cloud computing based mobile communication network data acquisition method can effectively exclude external interference and accurately collect data.

Conclusion

In this paper, a data acquisition method of mobile communication network based on cloud computing is proposed. The data eigenvector is extracted from the time-frequency distribution collected by the mobile communication network equipment to eliminate the interference from outside and complete the identification of the data eigenvectors. By using ADASYN algorithm, remove redundant information, accurate acquisition of mobile communication network data based on cloud computing. The experimental results show that the proposed method can effectively achieve the acquisition of mobile communication network data with high accuracy and efficiency.

References

[1] Polydoros A., Woo K.T. LPI detection of frequency hopping signal using autocorrelation techniques [J]. IEEE Journal on Selected Areas in Communications, 2009, 3(5): 714-726.

[2] Poisel R.A. Principles and Techniques of Modern Communication Interference [M]. Chen Ding Ding, et al. Beijing: Electronic Industry Press, 2005. (In Chinese)

[3] Xu R., Donald W.I. Survey of clustering algorithm [J]. IEEE Trans-actions on Neural Networks, 2005, 16(3): 645-678.

[5] Qian Yi, Wu Chuanhua. Research on Frequency Hopping Signal Detection and Parameter Estimation [J]. Space Electronics, 2007, 4 (2): 14-18. (In Chinese)

[6] Zhou Lin, Yang Shaoquan. Application of Time-Frequency Analysis Algorithm in Modulation Identification of Communication Signals [J]. A Space Electronic Warfare, 2002 (5): 40-45. (In Chinese)

[7] Guo Jun, Liu Heping, Liu Ping. Hall sensor based on high current detection applications [J]. Applied Technology, 2011 (5): 142-145. (In Chinese)

[8] Tan Tao. NET remote control excel in reserves management system application research [J]. Computer Applications and Software, 2007 (2): 43-44. (In Chinese)

[9] Yan Fei, Wang Sheng, Zhu Leiming. Intrusion Detection System Design Based on Data Fusion and Data Mining Technology [J]. Computer Engineering and Science, 2004, 26 (4): 15-18. (In Chinese)

[10] Li Hui. Network intrusion detection based on support vector machine [J]. Computer Research and Development, 2003, 40 (6): 799-807. (In Chinese)