Pattern updating device, updating method, and non-transitory computer-readable recording medium

Abstract

Provided is a pattern update device, a pattern update method, and a non-transitory computer-readable recording medium. The pattern update device extracts frequently occurring patterns from a plurality of pieces of path information, stores the extracted patterns as predetermined data structures, compares newly acquired path information with the stored predetermined data structures, and updates the stored predetermined data structures based on the comparison, thereby updating patterns that frequently occur in all pieces of path information including the newly acquired path information.

Description

Technical Field

[0001] This disclosure relates to mode update apparatus, mode update method, and non-transitory computer-readable recording medium. Background Technology

[0002] H. Cao et al., Mining frequent spatio-temporal sequential patterns, IEEE International Conference on Data Mining (ICDM), 2005 (hereinafter referred to as Cao et al.) and F. Giannotti et al., Efficient mining of temporally annotated sequences, SIAM International Conference on Data Mining (SDM), 2006 (hereinafter referred to as Giannotti et al.) disclosed the following method: taking multiple trips (road link sequences) as input, and extracting frequently occurring patterns through batch processing.

[0003] The methods described by Cao et al. and Giannotti et al. only provide a technique for batch processing multiple trips (road routes) as input to extract frequently occurring patterns. Pattern extraction is required every time a new trip (road route) is input. Furthermore, while the extracted patterns can often be aggregated using prefixes, simpler methods require matching based on the number of patterns multiplied by the pattern length. Here, a pattern refers to a group of specific travel points within a trip, and the pattern length refers to the number of specific travel points within a pattern. Summary of the Invention

[0004] The purpose of this disclosure is to provide a pattern updating apparatus, a pattern updating method, and a non-transient recording medium that reduces the processing load required for pattern extraction and updating.

[0005] The pattern update device of the first technical solution includes: a memory; and a processor connected to the memory. The processor extracts frequently occurring patterns from multiple path information, saves the extracted patterns as a predetermined data structure, compares newly acquired path information with the saved predetermined data structure, and updates the saved predetermined data structure based on the comparison, thereby updating the frequently occurring patterns in all path information including the newly acquired path information.

[0006] The pattern updating apparatus of the first technical solution extracts frequently occurring patterns from multiple path information sources, stores the extracted patterns as a predetermined data structure, compares newly acquired path information with the stored predetermined data structure, and updates the stored predetermined data structure based on the comparison, thereby updating frequently occurring patterns in all path information including newly acquired path information. According to the pattern updating apparatus of the first technical solution, by fitting frequently occurring patterns into a predetermined data structure and storing their information, and by comparing path information with the stored patterns and updating the patterns based on the comparison results, the processing load during pattern updates can be reduced. According to the pattern updating apparatus of the first technical solution, by comparing new path information based on the data structure, the comparison and frequency updates can be simplified.

[0007] In the pattern update device of the second technical solution, the processor is configured to extract path information that has exceeded a set number of times as frequently occurring patterns, wherein the set number of times is set to be less than the frequency of occurrence of the desired frequently occurring pattern.

[0008] According to the pattern update device of the second technical solution, frequent patterns are extracted by means of the minimum occurrence frequency of the pattern and the number of path information referenced, thus enabling pattern extraction with a margin of error.

[0009] In the pattern update device of the third technical solution, the processor is configured to adapt patterns that have become frequently occurring patterns through comparison to the data structure and save them.

[0010] According to the pattern updating device of the third technical solution, a pattern that has become a frequently occurring pattern through comparison can be used as a comparison object.

[0011] In the pattern update device of the fourth technical solution, the data structure is a data structure that aggregates common prefixes in a tree structure and maintains the occurrence count at the edges.

[0012] According to the pattern update device of the fourth technical solution, the data structure is simplified by aggregating common prefixes in a tree structure and maintaining the occurrence count at the edges.

[0013] In the pattern update method of the fifth technical solution, a processor extracts frequently occurring patterns from multiple path information, saves the extracted patterns as a predetermined data structure, compares newly acquired path information with the saved predetermined data structure, and updates the saved predetermined data structure based on the comparison, thereby updating the frequently occurring patterns in all path information containing the newly acquired path information.

[0014] In the pattern update method of the sixth technical solution, the processor extracts path information that has exceeded a set number of times as frequently occurring patterns, and the set number of times is set to be less than the frequency of occurrence of the desired frequently occurring pattern.

[0015] In the pattern update method of the seventh technical solution, the processor adapts the pattern that has become a frequently occurring pattern through comparison to the data structure and saves it.

[0016] In the pattern update method of the eighth technical solution, the data structure is a data structure that aggregates common prefixes in a tree structure and maintains the occurrence count at the edges.

[0017] The ninth technical solution provides a computer-readable non-transient recording medium containing a program that causes a computer to perform the following processing: extracting frequently occurring patterns from multiple path information, storing the extracted patterns as a predetermined data structure, comparing the stored predetermined data structure with newly acquired path information, and updating the stored predetermined data structure based on the comparison, thereby updating the frequently occurring patterns in all path information containing the newly acquired path information.

[0018] The tenth technical solution involves a computer-readable non-transient recording medium that extracts path information more than a set number of times as a frequently occurring pattern, wherein the set number of times is set to be less than the frequency of occurrence of the desired frequently occurring pattern.

[0019] The 11th technical solution uses a computer-readable non-transient recording medium to adapt and store frequently occurring patterns that have been re-established through comparison to the data structure.

[0020] In the computer-readable non-transient recording medium of the 12th technical solution, the data structure is a data structure that aggregates common prefixes in a tree structure and maintains the occurrence count at the edges.

[0021] The effects of the invention

[0022] According to this disclosure, a pattern updating apparatus, a pattern updating method, and a non-transient computer-readable recording medium can be provided, which extracts patterns of frequently occurring path information from the location information of vehicle driving data, efficiently matches new location information with the patterns when new location information is obtained, and updates the frequently occurring patterns, thereby reducing the processing load. Attached Figure Description

[0023] Figure 1 This is a diagram illustrating the configuration of an example of a pattern update device involved in an implementation of the disclosed technology.

[0024] Figure 2 This is a diagram illustrating an example of a data structure that represents the schema stored in the storage unit.

[0025] Figure 3 This is a diagram illustrating an example of comparison performed by the comparison section and update performed by the update section.

[0026] Figure 4 This is a diagram illustrating an example of comparison performed by the comparison section and update performed by the update section.

[0027] Figure 5 This is a diagram illustrating an example of comparison performed by the comparison section and update performed by the update section.

[0028] Figure 6 This is a diagram illustrating an example of comparison performed by the comparison section and update performed by the update section.

[0029] Figure 7 This is a diagram illustrating an example of comparison performed by the comparison section and update performed by the update section.

[0030] Figure 8 It is a flowchart representing the process of mode update handling performed by the mode update device. Detailed Implementation

[0031] Hereinafter, an example of an embodiment of the present disclosure will be described with reference to the accompanying drawings. Furthermore, the same or equivalent constituent elements and portions are given the same reference numerals in the drawings. Additionally, for ease of explanation, the scale of the drawings may sometimes be exaggerated and differ from the actual scale.

[0032] Figure 1 This diagram illustrates the configuration of the mode update device according to this embodiment. Figure 1The pattern update device 10 shown is a device that extracts frequently occurring trajectory patterns from the location information of vehicle driving data, efficiently matches new location information with the patterns, and updates the frequently occurring patterns.

[0033] The mode update device 10 has a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, and a storage 14, which are examples of hardware processors.

[0034] CPU 11 is a central processing unit that executes various programs and controls various components. Specifically, CPU 11 reads programs from ROM 12 or memory 14 and executes the programs using RAM 13 as its working area. CPU 11 performs control and various arithmetic operations on the aforementioned components according to the programs recorded in ROM 12 or memory 14. In this embodiment, a mode update program for performing mode update processing is stored in ROM 12 or memory 14.

[0035] ROM12 stores various programs and data. RAM13 serves as a working area to temporarily store programs or data. Storage 14 is composed of storage devices such as HDD (Hard Disk Drive), SSD (Solid State Drive), or flash memory, and stores various programs, including the operating system, and various data.

[0036] When executing the above-described pattern update procedure, the pattern update device 10 uses the aforementioned hardware resources to implement various functions. The functional configuration implemented by the pattern update device 10 will be described below.

[0037] like Figure 1 As shown, the mode update device 10 reads and executes the parsing program stored in the ROM 12 or the memory 14 through the CPU 11, thereby enabling the CPU 11 to function in such a way that it has an extraction unit 101, a storage unit 102, a comparison unit 103, and an update unit 104.

[0038] The extraction unit 101 extracts frequently occurring patterns from multiple pre-acquired path information sets. The path information can be stored, for example, in the storage device 14. For example, the extraction unit 101 can also extract path information that has been selected more than a set number of times as a frequently occurring pattern, where the set number is set to be less than the frequency of occurrence of the desired frequently occurring pattern.

[0039] In detail, the extraction unit 101 extracts frequently occurring patterns from the location information of multiple vehicle driving data. Location information, for example, refers to a sequence of GPS latitude and longitude (trip). By associating latitude and longitude with, for example, predefined discrete road routes, each trip is represented as a series of road routes as follows. In the example of trip 1, this means traveling from road A through road B, and further through road C. Each road route is also referred to as a node.

[0040] Itinerary 1: A→B→C

[0041] Itinerary 2: A→B→C→D

[0042] Itinerary 3: A→B→D→C→D

[0043] For the performance of such journeys, the extraction unit 101 extracts patterns of frequently occurring trajectories, for example, using the sequential pattern mining method proposed in references 1 and 2 below. At this time, the user specifies the minimum frequency m of the pattern occurrence in the total number of journeys M and the number N of journeys that match the extracted pattern in the comparison unit 103 (described later). The extraction unit 101 extracts only patterns that occur (m-N) or more times. Additionally, as a condition, the user can specify a ratio r (=(m-N) / M).

[0044] (Reference 1)

[0045] ·J.Pei et al., PrefixSpan: Mining sequential patterns efficiently by prefix-projected pattern growth. In Proceedings of International Conference onData Engineering (ICDE), 2001.

[0046] (Reference 2)

[0047] ·R.Srikant and R.Agrawal, Mining sequential patterns: generalizations and performance improvements. In Proceedings of the 5th International Conference on Extending Database Technology Advances in Database Technology (EDBT), 1996.

[0048] Furthermore, when comparing patterns with frequency m in the comparison unit 103, the reason for setting the number of occurrences of the patterns pre-extracted for creating the tree structure in the extraction unit 101 to (m-N) is to allow for leeway in the extraction of frequently occurring patterns. For example, if a frequently occurring pattern with frequency m=10 is desired, and the number of newly added strokes N is set to N=4, the tree structure is created using only strokes that initially appeared at least m-N=6 times in the extraction unit 101. By extracting in this way by the extraction unit 101, the comparison result of the new strokes with the pattern having a frequency of 6 shows that, with all N (=4) strokes being identical, a frequently occurring pattern with a frequency of 10 can be obtained.

[0049] The storage unit 102 stores the patterns extracted by the extraction unit 101 as a predetermined data structure. This predetermined data structure can be, for example, the data structures listed in references 1 and 2 above, or a data structure that aggregates common prefixes in a tree structure and maintains the frequency of occurrence at the edges. Furthermore, the storage unit 102 can also function as a memory 14.

[0050] Figure 2 This is a diagram illustrating an example of a data structure representing the mode stored in storage unit 102. Figure 2 These are examples saved for the following patterns.

[0051] Pattern 1: A→B (Frequency = 4)

[0052] Pattern 2: A→B→C (Frequency = 3)

[0053] Pattern 3: A→C (Frequency = 3)

[0054] Pattern 4: A→D (Frequency = 4)

[0055] Pattern 5: E→D (Frequency = 4)

[0056] Pattern 6: E→F (Frequency = 3)

[0057] The comparison unit 103 compares the newly acquired path information of the mode update device 10 with the predetermined data structure stored in the storage unit 102.

[0058] The updating unit 104 updates the predetermined data structure stored in the storage unit 102 based on the comparison of path information implemented by the comparison unit 103, thereby updating the patterns that frequently appear in all path information including the path information newly acquired by the pattern updating device 10. The updating unit 104 may also adapt patterns that have become frequently appearing patterns through the comparison of the comparison unit 103 to the predetermined data structure and store them in the storage unit 102.

[0059] The comparison performed by the comparison unit 103 and the update performed by the update unit 104 are repeated until the obtained path information is no longer available.

[0060] Figures 3-7 This diagram illustrates an example of comparison performed by the comparison unit 103 and update performed by the update unit 104. Figure 3 This is a comparative example when the route A→B→C has been newly obtained.

[0061] First, the starting point is node A; therefore, the comparison unit 103 first compares node A. Furthermore, as... Figure 3 As shown, the update unit 104 increases the occurrence count of node A from 5 to 6.

[0062] The next node is node B, therefore, the comparison unit 103 compares the path from node A to node B. Furthermore, as... Figure 4 As shown, the update unit 104 increases the number of occurrences of the path from node A to node B from 4 to 5.

[0063] The next node is node C, therefore, the comparison unit 103 compares the path from node B to node C. Furthermore, as... Figure 5 As shown, the update unit 104 increases the number of occurrences of the path from node A to node C via node B from 3 to 4.

[0064] The journey A→B→C can also be considered with node B as the starting point. First, since node B is the starting point, the comparison unit 103 first compares node B. Furthermore, as... Figure 6 As shown, the update unit 104 increases the occurrence count of node B from 4 to 5.

[0065] The next node is node C, therefore, the comparison unit 103 compares the path from node B to node C. Furthermore, as... Figure 7 As shown, the update unit 104 increases the number of occurrences of the path to node C from 3 to 4.

[0066] like Figures 3-7 As shown, the update unit 104 increases the occurrence count of the corresponding pattern's path by one while adding a new pattern that is at least as specified by the user as having a minimum frequency m or minimum proportion r.

[0067] If a new pattern that did not previously exist appears, the comparison unit 103 temporarily stops at a node along the path during comparison. For example, if the path A→Z→B occurs, it stops at node Z, which is a child node starting from node A, and updates the frequency of A→B when it reaches the preceding node B. In this case, if A→Z exceeds the frequency threshold, it is registered as a new pattern.

[0068] When the comparison unit 103 compares the A→B→C route and the update unit 104 updates the frequency, the comparison has two modes: A→B→C and B→C. However, the frequency of B→C is updated only once. In addition, the storage unit 102 can also distinguish between the B→C mode that starts from node A and the B→C mode that starts from node B and store them as other information.

[0069] When a new pattern such as route A→B→C→A appears, the storage unit 102 can also save the pattern because of the newly appearing endpoint node A.

[0070] The mode update device 10 has Figure 1 The configuration shown extracts frequently occurring trajectory patterns from the vehicle's location information (GPS, etc.) and efficiently matches new location information with these patterns, updating the frequently occurring patterns. The pattern update device 10 reduces the processing load by efficiently matching and updating frequently occurring patterns.

[0071] Next, the function of the mode update device 10 will be explained.

[0072] Figure 8 This is a flowchart illustrating the pattern update process implemented by the pattern update device 10. The pattern update program is read from the ROM 12 or the storage 14 by the CPU 11 and executed in the RAM 13, thereby performing the pattern update process.

[0073] CPU 11 extracts frequently occurring patterns from multiple pre-acquired path information sets (step S101). The path information can be stored, for example, in memory 14. For example, CPU 11 can also extract path information that has been extracted more than a set number of times as a frequently occurring pattern, where the set number is set to be less than the frequency of occurrence of the desired frequently occurring pattern.

[0074] Next, in step S101, CPU 11 saves the pattern extracted in step S101 as a predetermined data structure (step S102). The predetermined data structure can be, for example, the data structure listed in references 1 and 2 above, or a data structure that aggregates common prefixes in a tree structure and maintains the occurrence frequency at the edges.

[0075] Next, in step S102, CPU 11 begins processing the new trajectory. First, CPU 11 sets the number N of newly read patterns and sets the variable n to 0 (step S103).

[0076] Next, in step S103, CPU11 determines whether variable n is less than or equal to the number N of processes that match the extracted pattern (step S104).

[0077] If the result of step S104 is n≤N (step S104; yes), then the CPU11 will compare the pattern with the path information newly obtained by the pattern update device 10 according to the predetermined data structure saved in step S102 (step S105).

[0078] Next, in step S105, the CPU 11 updates the predetermined data structure saved in step S102 based on the comparison in step S105, thereby updating the patterns that frequently appear in all path information including the path information newly obtained by the pattern update device 10 (step S106).

[0079] Next, in step S106, CPU11 determines whether there is still a read process (step S107). If there is still a read process (step S107; yes), then CPU11 increments the variable n by one (step S108) and returns to the determination in step S104. If there is no read process (step S107; no), then CPU11 ends the processing of a sequence.

[0080] On the other hand, the result of step S104 is that if n > N (step S104; no), then CPU11 returns to step S101.

[0081] The mode update device 10 executes Figure 4 The illustrated series of processes extracts frequently occurring trajectory patterns from the vehicle's location information (GPS, etc.), efficiently matches new location information with these patterns, and updates the frequently occurring patterns. The pattern update device 10 reduces the processing load by efficiently matching and updating frequently occurring patterns.

[0082] Although a storage device 14 is provided inside the mode update device 10, this disclosure is not limited to this example.

[0083] Furthermore, the mode update process, in which the CPU reads and executes the software (program) in the above embodiments, can also be executed by various processors other than the CPU. Examples of processors in this case include FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices) whose circuit structure can be changed after manufacturing, and ASICs (Application Specific Integrated Circuits), which are dedicated circuits for processors with circuit structures specifically designed to execute specific processes. Moreover, the mode update process can be executed by one of these various processors, or by a combination of two or more processors of the same or different types (e.g., multiple FPGAs or a combination of a CPU and an FPGA). In more detail, the hardware construction of these various processors is a circuit in which circuit elements such as semiconductor components are combined.

[0084] Furthermore, while the above embodiments describe a technical solution where the pattern update processing program is pre-stored (installed) in a ROM or memory, this is not a limitation. The program can also be provided on a non-transitory recording medium such as a CD-ROM (Compact Disk Read Only Memory), DVD-ROM (Digital Versatile Disk Read Only Memory), or USB (Universal Serial Bus) memory. Alternatively, the program can be downloaded from an external device via a network.

Claims

1. A pattern update device, comprising: Memory; and The processor, which is connected to the memory, The processor is configured as follows: Multiple path information is obtained from the location information of the vehicle's driving data. Sequence pattern mining is used to extract frequently occurring patterns from multiple path information. The extracted patterns are saved as a predefined data structure. The newly obtained path information from the vehicle's driving data location information will be compared with the saved predetermined data structure. Based on the comparison, the predetermined data structure is updated, thereby updating the patterns that frequently occur in all path information including the newly acquired path information. Patterns that have become frequently occurring due to exceeding a threshold, as determined by the comparison, are adapted to the predetermined data structure and stored in the memory.

2. The mode update device according to claim 1, The processor is configured to extract path information that has occurred more than a set number of times as frequently occurring patterns, wherein the set number of times is set to be less than the frequency of occurrence of the desired frequently occurring pattern.

3. The mode update device according to claim 1, The data structure is a tree structure that aggregates common prefixes and maintains the occurrence count at the edges.

4. A schema update method, Through the processor, Multiple path information is obtained from the location information of the vehicle's driving data. Sequence pattern mining is used to extract frequently occurring patterns from multiple path information. The extracted patterns are saved as a predefined data structure. The newly obtained path information from the vehicle's driving data location information will be compared with the saved predetermined data structure. Based on the comparison, the predetermined data structure is updated, thereby updating the patterns that frequently occur in all path information including the newly acquired path information. Patterns that have become frequently occurring due to exceeding a threshold are stored in the predetermined data structure.

5. The mode update method according to claim 4, The processor extracts path information that has occurred more than a set number of times as frequently occurring patterns. The set number of times is set to be less than the desired frequency of occurrence of the frequently occurring pattern.

6. The mode update method according to claim 4, The data structure is a tree structure that aggregates common prefixes and maintains the occurrence count at the edges.

7. A computer-readable non-transitory recording medium containing a program that causes a computer to perform the following processing: Multiple path information is obtained from the location information of the vehicle's driving data. Sequence pattern mining is used to extract frequently occurring patterns from multiple path information. The extracted patterns are saved as a predefined data structure. The newly obtained path information from the vehicle's driving data location information will be compared with the saved predetermined data structure. Based on the comparison, the predetermined data structure is updated, thereby updating the patterns that frequently occur in all path information including the newly acquired path information. Patterns that have become frequently occurring due to exceeding a threshold are stored in the predetermined data structure.

8. The computer-readable non-transient recording medium according to claim 7, Path information that occurs more than a set number of times is extracted as frequently occurring patterns, where the set number of times is set to be less than the desired frequency of occurrence of the frequently occurring pattern.

9. The computer-readable non-transient recording medium according to claim 7, The data structure is a tree structure that aggregates common prefixes and maintains the occurrence count at the edges.

Images