A method for identifying information transmission link based on sequential correlation of time series data

Abstract

The application discloses a kind of information transmission link identification method based on sequential association of time series data, it includes: selecting two nodes to be associated from node table;Extract the time that each node in the two nodes to be associated receives the same information, the time sent to another node, node content, node name;According to the extracted information, the feature type identification of the two nodes to be associated is carried out;If the feature type of node content is digital type or string type, it is judged whether there is association between the two nodes to be associated;Until the association of all nodes in node table is judged, the time sequence that any two nodes in node table receives the same information is compared respectively, and finally the transmission link of all nodes to the same information is obtained.The application can improve identification accuracy.

Description

Technical Field

[0001] This invention relates to the field of communication network technology, and in particular to a method for identifying information transmission links based on sequential correlation of time-series data. Background Technology

[0002] With the development of communication network technology, communication networks have changed from point-to-point data transmission to multi-point-to-multi-point data transmission, greatly increasing the complexity of communication networks. In complex networks, even if the transmission network between nodes is planned in advance at the macro level, there is still the problem of unknown actual data transmission links at the data level. Moreover, in actual use, there may still be unplanned transmission links between nodes, making data traceability, multi-node data joint evaluation, and other data applications more difficult. Summary of the Invention

[0003] The purpose of this invention is to address the difficulties in tracing intermediate node data and evaluating data transmission efficiency between nodes caused by the dispersion of data at each node during data tracing and multi-node joint evaluation. This invention provides a method for identifying information transmission links based on the sequential association of time-series data. It extracts key descriptive features of each node's data and uses different feature matching methods for different feature types to identify information transmission links, thereby improving the accuracy of information transmission link identification. This method plays an important role in data applications such as data tracing and multi-node joint evaluation.

[0004] This invention discloses a method for identifying information transmission links based on sequential correlation of time-series data, comprising:

[0005] Step 1: Select two nodes to be associated from the node table; wherein, all nodes in the node table represent the transmission of the same information in chronological order;

[0006] Step 2: Extract the time when each of the two nodes that need to be associated receives the same information, the time when it is sent to the other node, the node content, and the node name; wherein, the node content includes the time when the content of the received same information is processed and the content obtained after processing;

[0007] Step 3: Based on the information extracted in Step 2, identify the feature types of the two nodes that need to be associated;

[0008] Step 4: If the feature type of the node content and node name is numeric or string, then determine whether there is a relationship between the two nodes that need to be associated.

[0009] Step 5: Repeat steps 1 to 4 until the correlation between all nodes in the node table is determined. Compare the order in which any two nodes in the node table receive the same information to obtain the transmission link of all nodes to the same information.

[0010] Further, step 4 includes:

[0011] Step 41: If the feature type of the node content and the node name is numeric, select one numeric data from the node content and the node name corresponding to the two nodes respectively;

[0012] Step 42: Calculate the difference between the two selected numeric data types and determine whether the difference meets the preset threshold. If it does not meet the threshold, it means that there is no association between the two nodes, and continue to execute Step 1. If it meets the threshold, continue to execute Step 41 and Step 42 until there is no unselected numeric data in the node content and node name corresponding to the two nodes. At this time, if all the differences between the two nodes meet the preset threshold, it means that there is an association between the two nodes.

[0013] Furthermore, step 4 also includes:

[0014] Step 401: If the feature type is string, select one string type data from the node content and node name corresponding to the two nodes respectively;

[0015] Step 402: Perform string splitting on the two selected string data types respectively to obtain two corresponding string arrays;

[0016] Step 403: Compare each character in one of the two string arrays with each character in the other string array. If each character in both string arrays matches, it means that there is a relationship between the two nodes corresponding to the two strings; otherwise, there is no relationship.

[0017] Further, step 402 includes:

[0018] The two strings are split into individual characters, special characters are removed, and all characters are arranged in order to form a string array, resulting in two string arrays.

[0019] Further, step 403 includes:

[0020] Step 4031: Check whether each of the two string arrays contains Arabic numerals;

[0021] Step 4032: Check whether the two string arrays contain Chinese characters or numbers respectively;

[0022] Step 4033: Determine whether the two string arrays match by checking whether the numbers, number of digits, and order of digits are consistent, and whether all other characters in the two string arrays match. Finally, determine whether there is a relationship between the two nodes corresponding to the two string arrays.

[0023] Further, step 4031 includes:

[0024] The regular expression matching method is used to check whether the two string arrays contain Arabic numerals. If Arabic numerals are found, their positions in the string arrays are recorded.

[0025] Further, step 4032 includes:

[0026] Check whether each of the two string arrays contains Chinese numerals. If it does, convert the Chinese numerals to Arabic numerals.

[0027] Further, step 4033 includes:

[0028] Step 40331: Check if the numbers, number of digits, and order of digits in the two string arrays are the same; if they are different, the two string arrays are unrelated; if they are the same, continue to step 40332.

[0029] Step 40332: Detect the similarity of the remaining characters between the two string arrays respectively. Match each character in the two string arrays and compare the unmatched characters with the stop word list to determine whether there is a relationship between the two string arrays, thereby determining whether there is a relationship between the two nodes corresponding to them.

[0030] Further, step 40332 includes:

[0031] Detect the similarity of the remaining characters between the two string arrays respectively. Match each character in the two string arrays and record the matching results of all characters. If the two characters can match, the matching result is recorded as 1; if the two characters cannot match, the matching result is recorded as 0.

[0032] If the number of matching characters is relatively small compared to the total number of characters, then the unmatched characters with a larger proportion are compared with the stop word list. If the unmatched characters are stop words in the stop word list, then there may be a relationship between the two string arrays and the two nodes corresponding to them; otherwise, the two string arrays are not related, and there is no relationship between the two nodes corresponding to them.

[0033] Further, step 5 includes:

[0034] Determine whether there are any nodes in the node table that need to be associated. If there are, continue to execute steps 1 to 4. If not, compare the reception time of the same information of all associated nodes, sort them according to the reception time, and form the final information transmission link.

[0035] Because of the adoption of the above technical solution, the present invention has the following advantages:

[0036] 1. Improved Recognition Accuracy. This invention leverages the potential correlations of multi-dimensional data features to infer node correlations from data-level relationships. Based on the temporal relationships of the data, it derives the transmission order of nodes, thereby identifying the information transmission chain. Compared to identifying information transmission chains from node transmission points, this data-level identification of information transmission chains, based on the actual data transmission between nodes, is more capable of uncovering transmission relationships that are difficult to obtain from nodes alone. Through practical engineering comparisons, the recognition accuracy is improved by more than 20%.

[0037] 2. The algorithm flow can be solidified. This method is not limited by the number of nodes and features processed, and nodes and key descriptive features can be added at any time according to technology and future development. However, the feature matching method remains universal, and the flow is more fixed.

[0038] 3. Reduces system costs and is suitable for widespread adoption. This invention mainly involves software algorithm updates and does not require hardware modifications, thus reducing system update costs. Furthermore, the algorithm flow is fixed; simply add parameter columns and increase the number of iterations to the existing algorithm. This simplifies modification and makes it suitable for widespread adoption. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments recorded in the embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0040] Figure 1 This is a flowchart illustrating an information transmission link identification method based on sequential association of time-series data, according to an embodiment of the present invention. Detailed Implementation

[0041] The present invention will be further described in conjunction with the accompanying drawings and embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art should fall within the protection scope of the present invention.

[0042] See Figure 1This invention provides an embodiment of an information transmission link identification method based on sequential correlation of time-series data, which includes:

[0043] First, the two nodes to be associated need to be selected from the node table, and the node data needs to be read. All nodes in the node table represent the transmission of the same information in chronological order.

[0044] Secondly, key descriptive features of the data are extracted. Each node's data contains multiple dimensions of descriptive features, some of which are not unique or will be discarded during transmission. Therefore, it is necessary to extract descriptive features that are both unique and transmissible. The specific extraction method involves comparing the data structures of each node to extract the unique and transmissible descriptive features of each node's data. The key descriptive features of each node's data include the time when the same information was received, the time when it was sent to another node, the node content, and the node name. The node content includes the time when the same information was processed and the resulting content.

[0045] Next, feature matching is performed. This mainly includes three steps:

[0046] The first step is feature type identification. Node content and node names have two data types: numeric and string. Different feature matching methods need to be selected for different types. Feature types can be identified by directly reading the feature data type or by regular expression matching.

[0047] The second step is to select an appropriate feature matching method. For numeric types, the difference between two sets of numbers can be calculated. A threshold M can be set based on human experience. The difference is compared to the threshold M. If the difference is below the threshold, it indicates a possible correlation between the two sets of data in that dimension. If the difference is above the threshold, it indicates no correlation between the two sets of data, and the feature matching step is skipped. For string types, due to human factors, the same thing may have multiple different names, such as different word order, different ways of representing numbers, or the use of special symbols. Directly comparing strings may lead to incorrect associations of descriptive features expressing the same thing. Therefore, intuitive fuzzy semantic matching is needed. The specific method is as follows:

[0048] First, the string is split into individual characters, special characters are removed, and all characters are arranged into an array in order. Then, regular expression matching is used to detect whether the string contains Arabic numerals. If Arabic numerals are found, their positions in the array are recorded. Next, the string is checked for Chinese numerals. If Chinese numerals are found, they are converted to Arabic numerals. The similarity of the numbers, number of digits, and order of the two sets of feature parameters is checked. If they are different, the two sets of data are unrelated. When the numbers, number of digits, and order of the two sets of feature parameters are all the same, the similarity of the remaining characters is checked. Each character in the two sets of feature parameters is matched, and the result of each character matching is recorded. A match is recorded as 1, and a mismatch as 0. The number of records with 1 is counted. If the count equals the number of characters with fewer characters, the unmatched characters in the feature parameters with more characters are queried and compared with the stop word list. If a stop word is found, it means that the two sets of data may be related in that dimension. Otherwise, it means that the two sets of data are not related, and the feature matching step is skipped.

[0049] The third step is to determine if there are any more descriptive features that need to be matched: traverse the feature table to see if there are any more descriptive features that need to be matched. If there are, return to the first step and continue feature matching. If not, it means that the two sets of data are related.

[0050] Next, determine if there are any more nodes that need to be associated: traverse the node table to check if there are any more node data that need to be associated. If so, return to step one and continue associating node data. If not, it means that the two sets of data are associated.

[0051] Finally, for all nodes with data associations, sort them according to the time when the nodes receive the data, arrange all nodes in chronological order to form the final information transmission link, which can be displayed using a network diagram.

[0052] Specifically, the steps include the following:

[0053] First, you need to select the node to be associated and read the node data.

[0054] Secondly, key descriptive features of the data are extracted. The data of each node contains descriptive features in multiple dimensions. Some of these features are not unique or will be discarded during transmission. Therefore, it is necessary to extract descriptive features that are both unique and transferable. The specific extraction method is simply to compare the data structures of each node and extract the descriptive features that are both unique and transferable for each node.

[0055] Next, feature matching is performed. This mainly includes three steps:

[0056] The first step is feature type identification. There are two data types for describing features: numeric and string. Different feature matching methods need to be selected for different types. Feature types can be identified by directly reading the feature data type or by regular expression matching.

[0057] The second step is to select an appropriate feature matching method. For numeric types, the difference between two sets of numbers can be calculated. A threshold M can be set based on human experience. The difference is compared to the threshold M. If the difference is below the threshold, it indicates a possible correlation between the two sets of data in that dimension. If the difference is above the threshold, it indicates no correlation between the two sets of data, and the feature matching step is skipped. For string types, due to human factors, the same thing may have multiple different names, such as different word order, different ways of representing numbers, or the use of special symbols. Directly comparing strings may lead to incorrect associations of descriptive features expressing the same thing. Therefore, intuitive fuzzy semantic matching is needed. The specific method is as follows:

[0058] First, the string is split into individual characters, special characters are removed, and all characters are arranged into an array in order. Then, regular expression matching is used to detect whether the string contains Arabic numerals. If Arabic numerals are found, their positions in the array are recorded. Next, the string is checked for Chinese numerals. If Chinese numerals are found, they are converted to Arabic numerals. The similarity of the numbers, number of digits, and order of the two sets of feature parameters is checked. If they are different, the two sets of data are unrelated. When the numbers, number of digits, and order of the two sets of feature parameters are all the same, the similarity of the remaining characters is checked. Each character in the two sets of feature parameters is matched, and the result of each character matching is recorded. A match is recorded as 1, and a mismatch as 0. The number of records with 1 is counted. If the count equals the number of characters with fewer characters, the unmatched characters in the feature parameters with more characters are queried and compared with the stop word list. If a stop word is found, it means that the two sets of data may be related in that dimension. Otherwise, it means that the two sets of data are not related, and the feature matching step is skipped.

[0059] The third step is to determine if there are any more descriptive features that need to be matched: traverse the feature table to see if there are any more descriptive features that need to be matched. If there are, return to the first step and continue feature matching. If not, it means that the two sets of data are related.

[0060] Next, determine if there are any more nodes that need to be associated: traverse the node table to check if there are any more node data that need to be associated. If so, return to step one and continue associating node data. If not, it means that the two sets of data are associated.

[0061] Finally, for all nodes with data associations, sort them according to the time when the nodes receive the data, arrange all nodes in chronological order to form the final information transmission link, which can be displayed using a network diagram.

[0062] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for identifying information transmission links based on sequential correlation of time-series data, characterized in that, include: Step 1: Select two nodes to be associated from the node table; wherein, all nodes in the node table represent the transmission of the same information in chronological order; Step 2: Extract the time when each of the two nodes that need to be associated receives the same information, the time when it is sent to the other node, the node content, and the node name; wherein, the node content includes the time when the content of the received same information is processed and the content obtained after processing; Step 3: Based on the information extracted in Step 2, identify the feature types of the two nodes that need to be associated; Step 4: If the feature type of the node content and node name is numeric or string, then determine whether there is a relationship between the two nodes that need to be associated. Step 5: Repeat steps 1 to 4 until the correlation between all nodes in the node table is determined. Compare the order in which any two nodes in the node table receive the same information to obtain the transmission link of all nodes to the same information. Step 4 includes: Step 41: If the feature type of the node content and the node name is numeric, select one numeric data from the node content and the node name corresponding to the two nodes respectively; Step 42: Calculate the difference between the two selected numeric data types and determine whether the difference meets the preset threshold. If it does not meet the threshold, it means that there is no association between the two nodes, and continue to execute Step 1. If it meets the threshold, continue to execute Step 41 and Step 42 until there is no unselected numeric data in the node content and node name corresponding to the two nodes. At this time, if all the differences between the two nodes meet the preset threshold, it means that there is an association between the two nodes. Step 4 also includes: Step 401: If the feature type is string, select one string type data from the node content and node name corresponding to the two nodes respectively; Step 402: Perform string splitting on the two selected string data types respectively to obtain two corresponding string arrays; Step 403: Compare each character in one of the two string arrays with each character in the other string array. If each character in both string arrays matches, it means that there is a relationship between the two nodes corresponding to the two strings; otherwise, there is no relationship.

2. The method according to claim 1, characterized in that, Step 402 includes: The two strings are split into individual characters, special characters are removed, and all characters are arranged in order to form a string array, resulting in two string arrays.

3. The method according to claim 1, characterized in that, Step 403 includes: Step 4031: Check whether each of the two string arrays contains Arabic numerals; Step 4032: Check whether the two string arrays contain Chinese characters or numbers respectively; Step 4033: Determine whether the two string arrays match by checking whether the numbers, number of digits, and order of digits are consistent, and whether all other characters in the two string arrays match. Finally, determine whether there is a relationship between the two nodes corresponding to the two string arrays.

4. The method according to claim 3, characterized in that, Step 4031 includes: The regular expression matching method is used to check whether the two string arrays contain Arabic numerals. If Arabic numerals are found, their positions in the string arrays are recorded.

5. The method according to claim 3, characterized in that, Step 4032 includes: Check whether each of the two string arrays contains Chinese numerals. If it does, convert the Chinese numerals to Arabic numerals.

6. The method according to claim 3, characterized in that, Step 4033 includes: Step 40331: Check if the numbers, number of digits, and order of digits in the two string arrays are the same; if they are different, the two string arrays are unrelated; if they are the same, continue to step 40332. Step 40332: Detect the similarity of the remaining characters between the two string arrays respectively. Match each character in the two string arrays and compare the unmatched characters with the stop word list to determine whether there is a relationship between the two string arrays, thereby determining whether there is a relationship between the two nodes corresponding to them.

7. The method according to claim 6, characterized in that, Step 40332 includes: Detect the similarity of the remaining characters between the two string arrays respectively. Match each character in the two string arrays and record the matching results of all characters. If the two characters can match, the matching result is recorded as 1; if the two characters cannot match, the matching result is recorded as 0. If the number of matching characters is relatively small compared to the total number of characters, then the unmatched characters with a larger proportion are compared with the stop word list. If the unmatched characters are stop words in the stop word list, then there may be a relationship between the two string arrays and the two nodes corresponding to them; otherwise, the two string arrays are not related, and there is no relationship between the two nodes corresponding to them.

8. The method according to claim 1, characterized in that, Step 5 includes: Determine whether there are any nodes in the node table that need to be associated. If there are, continue to execute steps 1 to 4. If not, compare the reception time of the same information of all associated nodes, sort them according to the reception time, and form the final information transmission link.

Images