Big data analysis method based on credential information

Abstract

The invention relates to the technical field of data analysis, and discloses a credential information-oriented big data analysis method, which comprises the steps of data acquisition, feature extraction, data analysis and result output. In the aspect of decision basis provision, feature association is analyzed through a behavior pattern matrix, and the influence of network performance on a user is evaluated; accurate support is provided for operators, developers and the like in decision making of marketing, network configuration, application optimization and the like. In the aspect of personalized service implementation, personalized requirements of users are known according to characteristics such as communication time distribution and application preference, and precise services are implemented by means of behavior pattern analysis results, such as customizing preferential activities for specific user groups, so that the loyalty and satisfaction of the users are improved. In a word, the method effectively fills up the defects of the traditional technology, powerfully promotes optimization and development of related services of the creative and creative industry, and has extremely important significance and value in application of big data analysis to the creative and creative field.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and particularly to a big data analysis method based on information and communication technology (ICT) information. Background Art

[0002] In the current era of rapid development of digital information, the ICT industry has emerged vigorously and been widely applied in many fields. With the high popularity of mobile terminal devices and the continuous emergence of various network applications, a large amount of ICT information data has been generated. For example, in the daily use of mobile terminal users, data such as call behavior, text message communication, application usage, and the performance status of the connected network continue to accumulate.

[0003] On the one hand, for enterprises and service providers, it has become extremely crucial to deeply understand the behavior patterns of target users. In a highly competitive market environment, accurately grasping users' communication habits, application preferences, and the distribution laws of these behaviors under different times and network environments can provide strong support for personalized marketing, service optimization, and product innovation. However, traditional data analysis methods often can only perform simple statistical analysis on a single type of data, and it is difficult to comprehensively and systematically integrate multi-source data and mine the complex correlation relationships therein, unable to meet the need for in-depth insight into user behavior.

[0004] On the other hand, as an important factor affecting user experience and business development, the evaluation of network performance and the quantitative analysis of its impact on user behavior also face challenges. Parameters such as network latency and data transmission rate in different network environments (such as 4G, 5G, and WIFI) change constantly, and there are no effective analysis means for how these changes specifically affect user communication activity, application usage duration, and other behaviors. For example, when the network is unstable, it is difficult for traditional analysis to give an accurate answer as to how much it will reduce users' willingness to use certain applications, or whether it will change the distribution of users' communication time.

[0005] In summary, in the context of big data of ICT information, there is an urgent need for a method that can comprehensively collect multi-source data, accurately extract features, and deeply analyze user behavior patterns and the impact of network performance, so as to fill the gaps in the existing technology and provide a solid data-driven foundation for the efficient development and optimization of related businesses in the ICT industry. The present invention is precisely a big data analysis method based on ICT information proposed based on such a need. Summary of the Invention

[0006] The purpose of the present invention is to provide a big data analysis method based on ICT information, which solves the technical problems proposed in the background art.

[0007] The purpose of the present invention can be achieved through the following technical solutions:

[0008] A big data analysis method based on information technology application innovation includes the following steps:

[0009] The first step, data collection:

[0010] Collect relevant target data of information technology application innovation; among them, the relevant target data includes: the operating parameters of mobile terminal devices, which include data transmission rate; the behavior data of target users, which include call duration, number of text messages, and application usage duration; network node data, which includes network latency time;

[0011] The second step, feature extraction:

[0012] Extract features from the target data, and obtain a network speed vector, a network stability coefficient, the communication activity of the target user, a communication time distribution vector, and their application preference degrees for each application;

[0013] Among them, the network speed vector is a vector representation form used to describe the data transmission rate of the target user in different network environments and the proportion of usage duration in each network environment; the network stability coefficient is a quantitative index to measure the stability of the network environment experienced by the target user; the communication activity of the target user is an index comprehensively considering the call and text message communication of the target user within a specified period; the communication activity of the target user is an index comprehensively considering the call and text message communication of the target user within a specified period;

[0014] The third step, data analysis:

[0015] Conduct target user behavior pattern analysis and network performance impact analysis based on the feature extraction results;

[0016] The fourth step, result output:

[0017] Display the data analysis results to relevant personnel.

[0018] As a further solution of the present invention: the feature extraction method is as follows:

[0019] StepB1: Extract communication activity

[0020] Extract the number of calls and the number of text messages of the target user within a specified period, and record them as n call and n sms ;

[0021] Then through:

[0022] F call =n call / T0

[0023] F sms =n sms / T0

[0024] Calculate the call frequency F of the target user within the specified period cal l and the text message frequency F sms ;

[0025] where T0 is the duration within the specified period;

[0026] After that, through:

[0027] A comm = α1×F cal l + α2×F sms

[0028] Calculate the communication activity A of the target user comm ;

[0029] where α1 and α2 are the corresponding preset weight coefficients;

[0030] Step B2: Extraction of communication time distribution vector

[0031] Divide the specified period into several standard time periods, and then divide each standard time period into several sub - time periods;

[0032] Within each standard time period, obtain the call duration of each call, and then calculate the sum of the call durations of all calls within each standard time period, and mark it as TS according to the time trend i , i = 1, 2,... n, n represents the number of all standard time periods within the specified period;

[0033] Then, within each sub - time period corresponding to each standard time period, obtain the call duration of each call, and then calculate the sum of the call durations of all calls within each sub - time period, and mark it as TS0 according to the time trend i，j , j = 1, 2,... m, m represents the number of all sub - time periods within the standard time period;

[0034] After that, let the value of j be 1, 2,... m in turn, and through:

[0035]

[0036] Calculate the call duration ratio TSB of the target user in each sub - time period j ;

[0037] Then, based on the call duration ratio TSB of the target user in each sub - time period j , form the corresponding communication time distribution vector of the target user [TSB1, TSB2,... TSB m .

[0038] Step B3: Application preference extraction

[0039] Obtain the application usage duration of the target user for various applications within a specified period, and mark it as YS k , k = 1, 2,..., p, where p is the number of application categories used by the target user;

[0040] Then calculate the sum of the application usage durations of all application categories, and mark it as YT;

[0041] After that, through Calculate the application preference Papp of the target user for each application k ;

[0042] Step B4: Network stability coefficient extraction

[0043] Within a specified period, based on the network latency times obtained by the target user at multiple collection time intervals, and form a latency time series {D1, D2,..., De};

[0044] where e is the number of collection time intervals;

[0045] Then through:

[0046]

[0047] Calculate the network stability coefficient S net ;

[0048] where r = 1, 2,..., e, r represents the rth collection time interval, and DP represents the average value of all network latency times;

[0049] Step B5: Network speed vector extraction

[0050] Extract the data transfer rates of the target user in different network environments, and mark them as CS 4G , CS 5G and CS WIFI , and the usage durations in different network environments, and mark them as SS 4G , SS 5G and SS WIFI ;

[0051] Then through:

[0052]

[0053] Successively calculate the usage duration ratios SB 4G , SB 5G and SBWIFI ;

[0054] Among them, different network environments include 4G network, 5G network and WIFI;

[0055] Subsequently, a network speed vector [(CS 4G , SB 4G ), (CS 5G SB 5G ), (CS WIFI SB WIFI )] is formed according to the data transmission rate and usage duration ratio of the target user in different network environments.

[0056] As a further solution of the present invention: the analysis method of the target user behavior pattern is as follows:

[0057] First, construct a target user behavior pattern matrix M;

[0058] Among them, the rows represent different target users, and the columns represent the characteristics corresponding to communication activity, application preference, communication time distribution vector and network speed vector;

[0059] Subsequently, by calculating the correlation coefficients between the characteristics of each column in the target user behavior pattern matrix, the association relationship between different characteristics is judged;

[0060] The calculation formula of the correlation coefficient is as follows:

[0061] Select two characteristics from the target user behavior pattern matrix and denote them as X g and Y g , where g = 1, 2,..., q, and q represents the number of target users in the target user behavior pattern matrix;

[0062] Subsequently, through:

[0063]

[0064] Calculate the correlation coefficient E between the characteristics of each column XY ;

[0065] In the formula, XP is the average value of the corresponding characteristic X g , and YP is the average value of the corresponding characteristic Y g ;

[0066] After that, compare the correlation coefficient E between the characteristics of each column XY with the corresponding preset correlation threshold of the associated characteristics:

[0067] If the correlation coefficient between the corresponding characteristics is greater than the correlation threshold of the associated characteristics, it means that the correlation between the corresponding characteristics is high.

[0068] As a further solution of the present invention, the network performance impact analysis method is as follows:

[0069] StepC2.1. Obtain the network stability coefficients of each user and mark them as S net，g ;

[0070] StepC2.2. Compare the network stability coefficient S of each user net，g with the preset network stability thresholds SY1 net and SY2 net and group the network stability levels of users according to the comparison results:

[0071] If S net，g > SY1 net , then classify the network stability level of the corresponding user into the low stability group;

[0072] If SY1 net ≥ S net，g > SY2 net , then classify the network stability level of the corresponding user into the medium stability group;

[0073] If S net，g ≤ SY2 net , then classify the network stability level of the corresponding user into the high stability group;

[0074] StepC2.3. In the low stability group and the high stability group, extract the application usage duration and communication activity of each group of users;

[0075] StepC2.4. Calculate the average values and variances of the corresponding characteristics of the application usage duration and communication activity of all users in the low stability group and the high stability group respectively;

[0076] At the same time, mark the average value of the application usage duration of all users in the low stability group as PU1, mark the variance of the application usage duration of all users in the low stability group as FU1, mark the average value of the application usage duration of all users in the high stability group as PU2, and mark the variance of the application usage duration of all users in the high stability group as FU2;

[0077] At the same time, mark the average value of the communication activity of all users in the low stability group as PH1, mark the variance of the communication activity of all users in the low stability group as FH1, mark the average value of the communication activity of all users in the high stability group as PH2, and mark the variance of the communication activity of all users in the high stability group as FH2;

[0078] StepC2.5. By:

[0079]

[0080] Calculate the evaluation index PG of the impact of network stability on the usage duration of applications respectively U and the evaluation index PG of the impact of network stability on communication activity H ;

[0081] StepC2.6. Compare the evaluation index PG of the impact of network stability on the usage duration of applications U with the preset usage duration evaluation threshold PGY U :

[0082] If PG U > PGY U , it means that the impact of network stability on the usage duration of applications is large;

[0083] Meanwhile, compare the evaluation index PG of the impact of network stability on communication activity H with the preset communication activity evaluation threshold PGY H :

[0084] If PG H > PGY H , it means that the impact of network stability on communication activity is large.

[0085] Advantages of the present invention:

[0086] In the present invention, by collecting relevant target data in multiple aspects such as the operation parameters of the mobile terminal device, the behavior data of the target user, and the network node data, and performing feature extraction, rich feature information such as network speed vectors, network stability coefficients, the communication activity of the target user, communication time distribution vectors, and their application preferences for each application can be obtained. This makes the description of the user and network conditions more comprehensive and detailed, providing a solid data foundation for subsequent accurate analysis.

[0087] In the present invention, a target user behavior pattern matrix is constructed. By calculating the correlation coefficients between the features in each column of the matrix, the correlation relationship between different features can be accurately judged. This helps to deeply understand the internal connection of the user's behavior characteristics in different aspects, such as knowing the degree of association between communication activity and application preference, etc., so as to more comprehensively grasp the user behavior pattern and provide strong support for personalized services, precision marketing, etc.

[0088] In the present invention, by comparing the network stability coefficient of a user with a preset threshold, the network stability levels of users are grouped. Then, the average values and variances of features such as the application usage duration and communication activity of users in different stability groups are calculated respectively, and evaluation indicators of the impact of network stability on the application usage duration and the impact on communication activity are obtained. In this way, the impact degree of network stability on user-related behaviors can be clearly understood, such as determining whether the impact of network stability on the application usage duration is large, etc., so as to take targeted measures such as network optimization to improve the user experience.

[0089] In the present invention, based on the above comprehensive and in-depth analysis results, whether it is the correlation analysis of user behavior patterns or the evaluation of the impact of network performance on user behavior, it can provide accurate and valuable decision-making basis for relevant personnel in formulating marketing strategies, optimizing network configurations, improving application functions, etc. For example, if it is found that the application preference degree of a certain type of application has a high correlation with communication activity, the operator can launch joint promotion activities for such user groups; if network stability has a great impact on the application usage duration, the network operator can focus on optimizing the network stability in the corresponding area.

[0090] In the present invention, through the quantitative analysis and evaluation of network stability coefficients, user behavior characteristics, etc., the problems or optimizable features existing in the network and services can be accurately located. For example, when it is found that the variances of the application usage duration and communication activity of users in the low-stability group are large, the network faults in the areas where these users are located can be checked or the network resource allocation can be optimized to improve the overall network service quality and thus improve user satisfaction.

[0091] In the present invention, with the help of information such as the communication time distribution vector and application preference degree of the target user obtained by feature extraction, the personalized needs and usage habits of each user can be deeply understood. For example, according to the proportion of call duration of the user in different time periods and the preference degree for various applications, more personalized service recommendations can be provided for the user, such as reducing push notifications during the user's call peak period, or recommending relevant value-added services according to the application types preferred by the user.

[0092] In the present invention, combined with the analysis results of the target user behavior patterns, such as the correlation relationships between various features, services can be implemented more precisely. For example, when a user group with high communication activity and high preference degree for a certain type of application is found, exclusive preferential activities can be carried out for this group or a better service experience can be provided to achieve precise services and improve user loyalty. BRIEF DESCRIPTION OF THE DRAWINGS

[0093] The present invention will be further described below with reference to the accompanying drawings.

[0094] Figure 1It is a schematic flowchart of a big data analysis method based on information and communication technology (ICT) information according to the present invention.

[0095] Figure 2 It is a schematic flowchart of feature extraction in a big data analysis method based on information and communication technology (ICT) information according to the present invention. Detailed implementation manners

[0096] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0097] Embodiment 1

[0098] Please refer to Figure 1 and Figure 2 As shown, the present invention is a big data analysis method based on information and communication technology (ICT) information, including the following steps:

[0099] The first step, data collection:

[0100] Collect relevant target data of information and communication technology (ICT) information;

[0101] The relevant target data includes but is not limited to:

[0102] The operating parameters of the mobile terminal device, which include the data transmission rate;

[0103] The behavior data of the target user, which includes the call duration, the number of text messages, and the application usage duration;

[0104] The network node data, which includes the network latency time;

[0105] The second step, feature extraction:

[0106] Extract features from the target data, and the specific method is as follows:

[0107] StepB cal 1: Extract the number of calls and the number of text messages of the target user within the specified period, and record them as n cal l and n sms ;

[0108] Then, through:

[0109] F call = n call / T0

[0110] F sms = n sms / T0

[0111] Calculate the call frequency F of the target user within the specified period call and the text message frequency F sms ;

[0112] Wherein, T0 is the duration within the specified period;

[0113] After that, through:

[0114] A comm = α1×F call +α2×F sms

[0115] Calculate the communication activity A of the target user comm ;

[0116] Wherein, α1 and α2 are the corresponding preset weight coefficients;

[0117] Step B2: Divide the specified period into several standard time periods, and then divide each standard time period into several sub-time periods;

[0118] Within each standard time period, obtain the call duration of each call, then calculate the sum of the call durations of all calls within each standard time period, and mark it as TS according to the time trend i , i = 1, 2,... n, n represents the number of all standard time periods within the specified period;

[0119] Then, within each corresponding sub-time period of each standard time period, obtain the call duration of each call, then calculate the sum of the call durations of all calls within each sub-time period, and mark it as TS0 according to the time trend i，j , j = 1, 2,... m, m represents the number of all sub-time periods within the standard time period;

[0120] After that, let the value of j be 1, 2,... m in sequence, and through:

[0121]

[0122] Calculate the call duration proportion TSB of the target user in each sub-time period j ;

[0123] Then, based on the call duration proportion TSB of the target user in each sub-time period j , form the corresponding communication time distribution vector [TSB1, TSB2,... TSB m of the target user

[0124] Step B3: Obtain the application usage duration of the target user using various applications within the specified period, and mark it as YS k, k = 1, 2, …… p, where p is the number of categories of applications used by the target user;

[0125] Then calculate the sum of the usage durations of all category applications and mark it as YT;

[0126] After that, through:

[0127]

[0128] Calculate the application preference degree Papp of the target user for each application k ;

[0129] StepB4: During the specified period, based on the network latency times obtained by the target user at multiple collection time periods, and form a latency time series {D1, D2, …… De};

[0130] where e is the number of collection time periods;

[0131] Then through:

[0132]

[0133] Calculate the network stability coefficient S net ;

[0134] where r = 1, 2, …… e, r represents the rth collection time period, and DP represents the average value of all network latency times;

[0135] StepB5: Extract the data transfer rates of the target user in different network environments and mark them as CS 4G 、CS 5G and CS WIFI , and the usage durations in different network environments and mark them as SS 4G 、SS 5G and SS WIFI ;

[0136] Then through:

[0137]

[0138] Calculate the usage duration ratios SB 4G 、SB 5G and SB WIFI ;

[0139] where different network environments include 4G network, 5G network and WIFI;

[0140] Subsequently, a network speed vector [(CS 4G , SB 4G ), (CS 5G SB 5G ), (CS WIFI SB WIFI )] is formed based on the data transmission rate and the usage duration ratio of the target user in different network environments;

[0141] Step 3: Data analysis:

[0142] Based on the feature extraction results, analyze the behavior patterns of the target user. The specific method is as follows:

[0143] First, construct a target user behavior pattern matrix M;

[0144] Among them, the rows represent different target users, and the columns represent the features corresponding to communication activity, application preference, communication time distribution vector, and network speed vector;

[0145] Subsequently, by calculating the correlation coefficients between the features in each column of the target user behavior pattern matrix, judge the association relationship between different features;

[0146] The calculation formula for the correlation coefficient is as follows:

[0147] Select two features from the target user behavior pattern matrix and denote them as X g and Y g , where g = 1, 2,... q, and q represents the number of target users in the target user behavior pattern matrix;

[0148] Subsequently, through:

[0149]

[0150] Calculate the correlation coefficient E XY between the features in each column;

[0151] In the formula, XP is the average value of the corresponding feature X g , and YP is the average value of the corresponding feature Y g ;

[0152] After that, compare the correlation coefficient E XY between the features in each column with the corresponding preset correlation threshold of the associated features:

[0153] If the correlation coefficient between the corresponding features is greater than the correlation threshold of the associated features, it means that the association between the corresponding features is high;

[0154] Example 1:

[0155] The correlation between communication activity and video application preference

[0156] If the correlation coefficient between communication activity and the preference degree for a certain video application is 0.65, while the preset correlation threshold for associated features is 0.5;

[0157] Since 0.65 > 0.5, this indicates that when the target user is relatively active in communication, there is a high probability that the user uses the video application;

[0158] The possible reason is that during the period when the user communicates frequently with others, the user is also more inclined to conduct richer social interactions or entertainment consumption through the video application;

[0159] Based on this result, the mobile operator can launch communication packages containing more video traffic or video application membership rights and interests for such users, and application developers can also conduct precise advertising placement or function optimization recommendations for such active users who prefer video applications;

[0160] Example 2:

[0161] Correlation between the communication time distribution vector and the network speed vector

[0162] If it is found that the proportion of the target user's call duration at night is relatively high, and during this time period, the proportion of the usage duration in the WIFI network environment is also relatively high. At the same time, there is a certain correlation between the effective speed corresponding to the WIFI network in the network speed vector and the communication activity during this time period. Assuming that the correlation coefficient is 0.52 and the threshold is 0.45;

[0163] It indicates that the user tends to use WIFI for communication activities at night, and the WIFI network speed affects the communication activity during this time period to a certain extent;

[0164] Based on this result, the operator can optimize the WIFI hotspot layout or network configuration during the night time period to improve the user experience; application developers can also push application functions or content updates that are suitable for the WIFI environment and have more usage scenarios at night according to the network environment and behavior characteristics of users during this time period;

[0165] Example 3:

[0166] Correlation between the application preference degree and the network stability coefficient

[0167] If the calculated correlation coefficient between the preference degree for a certain online game application and the network stability coefficient is 0.7, while the preset correlation threshold for associated features is 0.4;

[0168] This means that the degree of use of this online game application by the user is closely related to the network stability. When the network stability is good, the user is more willing to use this game application;

[0169] Based on this result, for game developers, it prompts them to optimize the performance of the game under different network stability conditions, or cooperate with the operator to ensure that some compensation mechanisms can be provided during network fluctuations, such as reducing the picture quality to maintain the smoothness of the game, etc.; for the operator, it can target the game player group, prioritize ensuring the network stability of their area, and launch a network acceleration service package specifically for game applications;

[0170] Step Four, Result Output:

[0171] Show the data analysis results to relevant personnel;

[0172] In this embodiment, a variety of target data related to Xinchuang information is collected through data acquisition, including mobile terminal device operation parameters, target user behavior data, network node data, etc. Then detailed feature extraction is carried out, covering various features such as communication activity, communication time distribution vector, application preference degree, network stability coefficient, network speed vector, etc. Based on the results of these feature extractions, the behavior patterns of different target users can be analyzed comprehensively and deeply; based on the correlation results between different features obtained from the analysis of the behavior patterns of target users, mobile operators can launch communication packages that better meet the needs of specific user groups, such as launching packages containing more video traffic or video application membership rights for users who are communication active and prefer video applications. Application developers can also carry out precise advertising placement or function optimization recommendations. For example, according to the network environment and behavior characteristics of users at different times, suitable application functions or content updates are pushed, so as to improve the user experience and achieve precise marketing and service optimization.

[0173] Embodiment Two

[0174] As Embodiment Two of the present invention, when this application is specifically implemented, compared with Embodiment One, the technical solution of this embodiment is only different from that of Embodiment One in that in the data analysis step, network performance impact analysis is also carried out based on the results of feature extraction, and the specific method is as follows:

[0175] StepC2.1: Obtain the network stability coefficient of each user and mark it as S net，g ;

[0176] StepC2.2: Compare the network stability coefficient S of each user net，g with the preset network stability thresholds SY1 net and SY2 net and group the network stability levels of users according to the comparison results:

[0177] If S net，g > SY1 net, the network stability level of the corresponding user will be classified into the low - stability group;

[0178] If SY1 net ≥S net，g >SY2 net , the network stability level of the corresponding user will be classified into the medium - stability group;

[0179] If S net，g ≤SY2 net , the network stability level of the corresponding user will be classified into the high - stability group;

[0180] StepC2.3. Extract the application usage duration and communication activity of each user in the low - stability group and the high - stability group;

[0181] StepC2.4. Calculate the average value and variance of the corresponding features of the application usage duration and communication activity of all users in the low - stability group and the high - stability group respectively;

[0182] At the same time, mark the average value of the application usage duration of all users in the low - stability group as PU1, the variance of the application usage duration of all users in the low - stability group as FU, mark the average value of the application usage duration of all users in the high - stability group as PU2, and the variance of the application usage duration of all users in the high - stability group as FU2;

[0183] At the same time, mark the average value of the communication activity of all users in the low - stability group as PH1, the variance of the communication activity of all users in the low - stability group as FH1, mark the average value of the communication activity of all users in the high - stability group as PH2, and the variance of the communication activity of all users in the high - stability group as FH2;

[0184] StepC2.5. Through:

[0185]

[0186] Calculate the evaluation index PG of the impact of network stability on application usage duration U and the evaluation index PG of the impact of network stability on communication activity H ;

[0187] StepC2.6. Compare the evaluation index PG of the impact of network stability on application usage duration U with the preset usage - duration evaluation threshold PGY U :

[0188] If PG U >PGY U , it means that the impact of network stability on application usage duration is large;

[0189] Meanwhile, the evaluation index PG for the impact of network stability on communication activity H is compared with the preset communication activity evaluation threshold PGY H as follows:

[0190] If PG H > PGY H , it indicates that the impact of network stability on communication activity is significant.

[0191] Based on Embodiment 1, Embodiment 2 further conducts an analysis of the impact of network performance according to the feature extraction results. By comparing the network stability coefficient of the user with the preset threshold, the network stability levels of the users are grouped. Then, features such as the application usage duration and communication activity of users in different groups are extracted, and their averages and variances are calculated. Furthermore, the evaluation index for the impact of network stability on the application usage duration and the evaluation index for the impact on communication activity are obtained. Through the above analysis, the impact degree of network stability on the application usage duration and communication activity can be clarified. If the evaluation index is greater than the corresponding preset threshold, it indicates that the impact of network stability on it is significant. This provides clear data support for relevant parties such as operators, helping them take targeted measures to optimize network configuration and ensure network stability, so as to better meet the needs of users in application usage and communication activities and improve the overall service quality.

[0192] Embodiment 3

[0193] As Embodiment 3 of the present invention, in the specific implementation of this application, compared with Embodiment 1 and Embodiment 2, the technical solution of this embodiment lies in combining the solutions of the above Embodiment 1 and Embodiment 2 for implementation.

[0194] Embodiment 3 combines the solutions of Embodiment 1 and Embodiment 2 for implementation, which can not only comprehensively and deeply understand the user behavior pattern, achieve precise marketing and service optimization, but also deeply evaluate the impact of network performance and provide a basis for network optimization. Combining the advantages of the previous two embodiments, it can analyze and process the data related to information and communication technology from multiple angles, so as to more comprehensively provide valuable information for relevant entities such as mobile operators and application developers, so that they can adopt more effective strategies to improve the user experience, optimize services and ensure network performance, etc.

[0195] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by collecting a large amount of data for software simulation to obtain a formula closest to the actual situation. The preset parameters and threshold selection in the formulas are set by those skilled in the art according to the actual situation.

[0196] As described above, it is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claimed rights.

Claims

1. A big data analysis method based on information and communication technology innovation-oriented information, characterized in that, It includes the following steps: Step 1: Collect relevant target data of information technology application innovation. Among them, the relevant target data includes: the operating parameters of mobile terminal devices, which include data transmission rate; the behavior data of target users, which includes call duration, number of text messages, and application usage duration; network node data, which includes network latency time. Step 2: Extract features from the target data, and obtain features corresponding to the network speed vector, network stability coefficient, communication activity of the target user, communication time distribution vector, and application preference degree for each application. Step 3: Analyze the behavior patterns of target users and the impact on network performance based on the feature extraction results, and then display the analysis results to relevant personnel.

2. The big data analysis method based on information and communication technology (ICT) information according to claim 1, wherein The extraction method of communication activity in feature extraction is as follows: Extract the number of calls and the number of text messages of the target user within the specified period, and record them as n call and n sms ; Then through: F call = n call / T0 F sms = n sms / T0 Calculate the call frequency F of the target user within the specified period call and the text message frequency F sms ; In the formula, T0 is the duration within a specified period. After that through: A comm = α1 × F call + α2 × F sms Calculate the communication activity A of the target user comm ; In the formula, α1 and α2 are preset weight coefficients.

3. A big data analysis method based on information technology application innovation information according to claim 2, characterized in that, The extraction method of the communication time distribution vector in feature extraction is as follows: Divide the specified period into several standard time periods, and then divide each standard time period into several sub-time periods. During each standard time period, obtain the call duration of each call, then calculate the sum of the call durations of all calls within each standard time period, and mark it as TS according to the time trend i , where i = 1, 2,... n, and n represents the number of all standard time periods within the specified cycle; Then, in each sub - time period corresponding to each standard time period, obtain the call duration of each call. Subsequently, calculate the sum of the call durations of all calls in each sub - time period, and mark it as TS0 according to the time trend. i，j , where j = 1, 2, …… m, and m represents the number of all sub - time periods within the standard time period; After that, let the value of j be 1, 2,..., m in sequence, and through: Calculate the call duration ratio TSB of the target user in each sub-time period j ; Subsequently, according to the proportion of call duration TSB of the target user in each sub-time period j , a communication time distribution vector [TSB1, TSB2,... TSB m of the corresponding target user is formed.

4. A big data analysis method based on information and communication technology (ICT) innovation-oriented information according to claim 3, characterized in that, The extraction method of application preference degree in feature extraction is as follows: Obtain the application usage duration of the target user for various applications within a specified period, and mark it as YS k , where k = 1, 2,..., p, and p is the number of categories of applications used by the target user; Then calculate the sum of the application usage durations of all categories of applications, and mark it as YT. After that through: Calculate the application preference degree Papp of the target user for each application k .

5. A big data analysis method based on information and communication technology (ICT) innovation-oriented information according to claim 4, characterized in that The extraction method of the network stability coefficient in feature extraction is as follows: Within the specified period, based on the network latency times obtained by the target user in multiple collection time periods, and form a latency time series {D1, D2,..., De}. Among them, e is the number of collection time periods. Then through: Calculate the network stability coefficient S net ; Among them, r = 1, 2,..., e, and r represents the rth collection time period.

6. A big data analysis method based on information creation-oriented information according to claim 5, characterized in that, The extraction method of the network speed vector in feature extraction is as follows: Extract the data transfer rates of the target users in different network environments and label them as CS 4G , CS 5G and CS WIFI , as well as the usage durations in different network environments, and label them as SS 4G , SS 5G and SS WIFI ; Then through: Calculate the usage duration ratios SB of the corresponding target users in different network environments in sequence 4G 、SB 5G and SB WIFI ; Among them, different network environments include 4G network, 5G network, and WIFI. Subsequently, a network speed vector [(CS 4G , SB 4G ), (CS 5G SB 5G ), (CS WIFI SB WIFI )] is formed based on the data transmission rate and usage duration ratio of the target user in different network environments.

7. A big data analysis method based on information creation-oriented information according to claim 6, characterized in that The analysis method of target user behavior patterns is as follows: First, construct a target user behavior pattern matrix M. Among them, the rows represent different target users, and the columns represent the features corresponding to communication activity, application preference degree, communication time distribution vector, and network speed vector. Then, judge the correlation relationship between different features by calculating the correlation coefficients between the features in each column of the target user behavior pattern matrix. The calculation formula of the correlation coefficient is as follows: From the target user behavior pattern matrix, select two features and denote them as X g and Y g , where g = 1, 2, …… q, and q represents the number of target users in the target user behavior pattern matrix; Then through: Calculate the correlation coefficient E between the features of each column XY ; Afterwards, the correlation coefficient E between each column of features XY is compared with the corresponding preset correlation threshold of associated features: If the correlation coefficient between the corresponding features is greater than the correlation threshold of associated features, it means that the correlation between the corresponding features is high.

8. A big data analysis method based on information and communication technology (ICT) information according to claim 6, characterized in that The analysis method of the impact on network performance is as follows: StepC2.

1. Obtain the network stability coefficient of each user and mark it as S net，g ; Step C2.2: Take the network stability coefficient S of each user net，g and compare it with the preset network stability thresholds SY1 net and SY2 net to group the network stability levels of the users according to the comparison results, and obtain the low-stability group, medium-stability group, and high-stability group: StepC2.3: In the low-stability group and the high-stability group, extract the application usage durations of each group of users. StepC2.4: Calculate the average value and variance of the features corresponding to the application usage durations of all users in the low-stability group and the high-stability group respectively. At the same time, mark the average value of the application usage durations of all users in the low-stability group as PU1, mark the variance of the application usage durations of all users in the low-stability group as FU1, mark the average value of the application usage durations of all users in the high-stability group as PU2, and mark the variance of the application usage durations of all users in the high-stability group as FU2. StepC2.5: Through: Calculate the evaluation index PG of the impact of network stability on the usage duration of the application U ; StepC2.

6. Evaluate the impact of network stability on the usage duration of the application, with the evaluation metric PG U compared with the preset usage duration evaluation threshold PGY U as follows: If PG U > PGY U , it means that the network stability has a great impact on the usage duration of the application.

9. A big data analysis method based on information and communication technology (ICT) - oriented information according to claim 8, characterized in that The comparison method in Step C2.2 is as follows: If S net，g > SY1 net , then the network stability level of the corresponding user will be classified into the low stability group; If SY1 net ≥S net，g >SY2 net , then the network stability level of the corresponding user will be classified into the medium stability group; If S net，g ≤ SY2 net , then the network stability level of the corresponding user will be classified into the high stability group.

10. A big data analysis method based on information creation-oriented information according to claim 8, characterized in that, The network performance impact analysis method is also as follows: First, in the low-stability group and the high-stability group, extract the communication activity of each user; Then, calculate the average value and variance of the corresponding features of the communication activity of all users in the low-stability group and the high-stability group respectively; At the same time, mark the average value of the communication activity of all users in the low-stability group as PH1, mark the variance of the communication activity of all users in the low-stability group as FH1, mark the average value of the communication activity of all users in the high-stability group as PH2, and mark the variance of the communication activity of all users in the high-stability group as FH2; Then through: Calculate the evaluation index PG of the impact of network stability on communication activity H ; Afterwards, the evaluation index PG of the impact of network stability on communication activity H is compared with the preset communication activity evaluation threshold PGY H as follows: If PG H > PGY H , it means that the network stability has a great impact on the communication activity.

Images