JP7168826B2 - Data integration support device, data integration support method, and data integration support program - Google Patents

Description

The present invention relates to a data integration support device, a data integration support method, and a data integration support program.

It is known that system integration or the like requires database integration processing (see, for example, Patent Document 1).

JP 2016-136354 A

By the way, in database integration processing, when the integration target and the integration destination are structured data, a technique called schema matching is used to match two schemas. For example, if the integration target and the integration target are structured data in table format, the individual attribute name data set of the integration target is set as the matching target schema, and the individual attribute name data set of the integration target is set as the matching target schema. matching is done.

However, the expression of attribute name data (hereinafter simply referred to as attribute name) may not be unified. For example, "company" may be defined as an attribute name to be integrated, and "company name" may be defined as an attribute name to be integrated. In this case, if each attribute value data (hereinafter simply referred to as attribute value) represents the name of a specific company, these two attribute names are syntactically inconsistent but semantically A person can determine that they are synonymous and can combine them into two attribute names.

However, when schema matching is performed mechanically by software processing, etc., as mentioned above, the attribute name to be integrated and the attribute name to be integrated may not syntactically match. It is difficult to determine whether names can be merged. On the other hand, when attempting to mechanically determine the meaning of an attribute name by extracting attribute values, there is a problem that a large amount of calculation is required due to the enormous number of attribute values.

Therefore, one aspect aims to support efficient data integration.

In one embodiment, the data integration support device includes first structured data including a first attribute name and a first attribute value corresponding to the first attribute name, a plurality of second attribute names, and the a storage unit for storing second structured data including a second attribute value corresponding to each of a plurality of second attribute names; first document information obtained by aggregating the first attribute values; based on the first extraction method of extracting together with the important words the syntactic feature amount representing the feature amount related to the syntax of the word, the second document information is generated by aggregating the attribute values of each attribute name. A first syntactic feature quantity of a word included in the first attribute value is extracted from one piece of document information, and a second syntactic feature quantity of a word included in the second attribute value is extracted from the second document information. is extracted, a specific important word is extracted together with the extraction of the first syntactic feature amount and the second syntactic feature amount, and a semantic feature amount representing a feature amount related to the meaning of the word is extracted. Based on a second extraction method, extracting a specific semantic feature amount corresponding to the specific important word from the specific important word, and applying the first syntactic feature amount to the specific semantic feature amount generating a second synthetic feature by weighting the first synthetic feature and the specific semantic feature with the second syntactic feature, and combining the first synthetic feature and the second syntactic feature; A process of outputting at least one of a first degree of similarity representing a degree of similarity of combined feature quantities and a specific second attribute name indicating a destination of integration of the first attribute name determined according to the first degree of similarity. and a processing unit that executes

It can support efficient data integration.

FIG. 1 is an example of a terminal device. FIG. 2 is an example of the hardware configuration of a terminal device. FIG. 3 is an example of a block diagram of a terminal device. FIG. 4 is a flowchart (part 1) showing an example of processing executed by the terminal device. FIG. 5(a) is an example of the first structured data. FIG. 5(b) is an example of the second structured data. FIG. 6 is a diagram for explaining an example of generating document information from the first structured data. FIG. 7 is a diagram for explaining an example of generating document information from the second structured data. FIG. 8 is a diagram for explaining an example of classification of document information. FIG. 9 is a diagram for explaining an example of data cleansing. FIG. 10 is a diagram for explaining an example of morphological analysis. FIG. 11 is a diagram for explaining an example of extraction of syntactic features and important words. FIG. 12 is a diagram for explaining an example of extraction of semantic features. FIG. 13 is a diagram for explaining an example of generating a synthetic feature amount. FIG. 14 is a diagram for explaining a matching example when the attribute value is a character. FIG. 15 is a flowchart (part 2) showing an example of processing executed by the terminal device. FIG. 16 is a diagram for explaining an example of attribute name extraction. FIG. 17 is another example of morphological analysis. FIG. 18 is a diagram for explaining another example of extraction of semantic features. FIG. 19 is a diagram for explaining an example of generating an average feature amount. FIG. 20 is a diagram for explaining a matching example when attribute values are numerical values. FIG. 21 is a diagram for explaining another matching example when attribute values are numerical values. FIG. 22 is an example of the matching list screen. FIG. 23 is a diagram briefly explaining the advantages of the first embodiment. FIG. 24 is a diagram for explaining a matching example according to the comparative example. FIG. 25 is a diagram for explaining a matching example according to the second embodiment. FIG. 26 is an example of a data integration support system.

Hereinafter, the form for carrying out this case will be described with reference to the drawings.

(First embodiment)
FIG. 1 is an example of a terminal device 100. As shown in FIG. The terminal device 100 is an example of a data integration support device. Although FIG. 1 shows a personal computer (PC) as an example of the terminal device 100, it may be a smart device such as a tablet terminal. The terminal device 100 includes a keyboard and pointing device (hereinafter simply referred to as keyboard) 100F. The terminal device 100 has a display 100G. The display 100G may be a liquid crystal display or an organic electro-luminescence (EL) display.

The display 100G displays various screens. Although the details will be described later, for example, the display 100G displays the matching list screen 10 including attribute names to be integrated, candidates for attribute names to be integrated with the attribute names, and ranking of the candidates. A user (for example, a data scientist) using the terminal device 100 confirms the ranking of the candidates included in the matching list screen 10, and operates the keyboard 100F to select one of the candidates for the attribute name to be integrated. to select. As a result, the terminal device 100 integrates the attribute names to be integrated into the selected candidates. By using the information presented on the matching list screen 10 in this way, the user can easily determine the attribute name for data integration, and the efficiency of data integration can be improved.

Next, a hardware configuration of the terminal device 100 will be described with reference to FIG.

FIG. 2 shows an example of the hardware configuration of the terminal device 100. As shown in FIG. As shown in FIG. 2, the terminal device 100 includes at least a Central Processing Unit (CPU) 100A as a hardware processor, a Random Access Memory (RAM) 100B, a Read Only Memory (ROM) 100C, and a network I/F (interface). Includes 100D. Moreover, as described above, the terminal device 100 also includes a keyboard 100F and a display 100G.

Furthermore, the terminal device 100 may include at least one of a hard disk drive (HDD) 100E, an input/output I/F 100H, a drive device 100I, and a short-range wireless communication circuit 100J, if necessary. The CPU 100A to the short-range wireless communication circuit 100J are connected to each other by an internal bus 100K. That is, the terminal device 100 can be implemented by a computer. A Micro Processing Unit (MPU) may be used as a hardware processor instead of the CPU 100A.

A semiconductor memory 730 is connected to the input/output I/F 100H. Examples of the semiconductor memory 730 include Universal Serial Bus (USB) memory and flash memory. The input/output I/F 100H reads programs and data stored in the semiconductor memory 730 . The input/output I/F 100H has, for example, a USB port. A portable recording medium 740 is inserted into the drive device 100I. Examples of the portable recording medium 740 include removable discs such as Compact Disc (CD)-ROM and Digital Versatile Disc (DVD). The drive device 100I reads programs and data recorded on the portable recording medium 740 . The short-range wireless communication circuit 100J is an electric circuit or electronic circuit that realizes short-range wireless communication such as Wi-Fi (registered trademark) or Bluetooth (registered trademark). An antenna 100J' is connected to the short-range wireless communication circuit 100J. A CPU that implements a communication function may be used instead of the short-range wireless communication circuit 100J. The network I/F 100D has, for example, a Local Area Network (LAN) port.

The programs stored in the ROM 100C and the HDD 100E are temporarily stored in the above-described RAM 100B by the CPU 100A. The program recorded on the portable recording medium 740 is temporarily stored in the RAM 100B by the CPU 100A. By executing the stored programs, the CPU 100A implements various functions described later, and also executes various processes described later. Note that the program may be in accordance with the flow chart described later.

Next, the functional configuration of the terminal device 100 will be described with reference to FIG.

FIG. 3 is an example of a block diagram of the terminal device 100. As shown in FIG. FIG. 3 shows main functions of the terminal device 100 . As shown in FIG. 3, the terminal device 100 includes a storage section 110, a processing section 120, an input section 130, and a display section 140. FIG. The storage unit 110 can be realized by the above-described RAM 100B and HDD 100E. The processing unit 120 can be realized by the CPU 100A described above. The input unit 130 can be implemented by the keyboard 100F described above. The display unit 140 can be realized by the display 100G described above. Therefore, the storage section 110, the processing section 120, the input section 130, and the display section 140 are connected to each other.

Here, the storage unit 110 includes a data storage unit 111, a document storage unit 112, and a dictionary storage unit 113 as constituent elements. The storage unit 110 also includes a syntactic feature storage unit 114, a vector storage unit 115, a semantic feature storage unit 116, and a composite feature storage unit 117 as components. On the other hand, the processing unit 120 includes a document generation unit 121, a cleansing unit 122, and an analysis unit 123 as constituent elements. The processing unit 120 also includes a first extraction unit 124, a second extraction unit 125, a feature synthesis unit 126, and a similarity output unit 127 as components.

Each component of the processing unit 120 accesses at least one of the components of the storage unit 110 to perform various processes. For example, the document generation unit 121 accesses the data storage unit 111 and acquires structured data stored in the data storage unit 111 . When the document generation unit 121 acquires the structured data, it generates document information in which attribute values included in the structured data are aggregated, and stores the document information in the document storage unit 112 . Other components will be described in detail when the operation of the terminal device 100 is described.

Next, operations of the terminal device 100 will be described with reference to FIGS. 4 to 21. FIG.

First, the data storage unit 111 stores two pieces of structured data (step S101). More specifically, as shown in FIGS. 5A and 5B, the data storage unit 111 stores first structured data D1 and second structured data D2 as two types of structured data. Both the first structured data D1 and the second structured data D2 are electronic files whose data structure is defined in a table format. In the first embodiment, the first structured data D1 corresponds to the integration target, and the second structured data D2 corresponds to the integration target.

The two structured data stored in the data storage unit 111 may be electronic files whose data structures are defined in the comma-separated values (CSV) format, or electronic files whose data structures are defined in the fixed-length format. It can be a file. Since the data structure of the table format, CSV format, etc. is completely defined, structured data can be called completely structured data. On the other hand, structured data does not include unstructured data, such as image files and audio files, whose data structure is not defined at all. The two structured data described above need only be stored in the data storage unit 111 before subsequent processing is executed, and the storage path of the two structured data is not particularly limited.

As shown in FIG. 5A, the first structured data D1 includes attribute names and attribute values corresponding to the attribute names. Specifically, the first structured data D1 includes "company", "transfer date", etc. as attribute names. Further, the first structured data D1 includes "Izumo Metals Co., Ltd." and "February 2, 2012" as attribute values. Incidentally, in the first embodiment, as shown in FIG. The first structured data D1 may include one attribute name and one attribute value corresponding to the attribute name. That is, the number of attribute names included in the first structured data D1 corresponding to the integration target is not particularly limited.

On the other hand, as shown in FIG. 5B, the second structured data D2 includes a plurality of attribute names and attribute values corresponding to each of the plurality of attribute names. Specifically, the second structured data D2 includes "position", "birthday", etc. as attribute names. Further, the second structured data D2 includes "researcher" and "March 12, 1987" as attribute values. As described above, the attribute name included in the second structured data D2 corresponding to the integration destination does not syntactically match the attribute name included in the first structured data D1 corresponding to the integration target, but when the attribute value is referred to , there are attribute names that are semantically synonymous. For example, "company name" included in the second structured data D2 and "company" included in the first structured data D1 do not match syntactically because the forms of the words do not match. and "Izumo Kinzoku Co., Ltd." included in the first structured data D1 are assumed to represent the name of a specific company, so "company name" and "company" are semantically synonymous. Such a relationship is the same for “birthday” and “date of birth”. Also, although not shown, such a relationship is the same for "air conditioner" and "air conditioner".

When the process of step S101 is completed, the document generation unit 121 then generates document information (step S102). Specifically, the document generation unit 121 accesses the data storage unit 111 and acquires the first structured data D1 and the second structured data D2 stored in the data storage unit 111 . When the document generation unit 121 acquires the first structured data D1 and the second structured data D2, as shown in FIG. Document information is generated column by column (ie, column by column). Then, the document generation unit 121 associates each piece of document information with an identifier for identifying the document information, and holds the document information. Next, as shown in FIG. 7, the attribute values included in the second structured data D2 are aggregated and separated by spaces to generate document information for each column. Then, the document generation unit 121 associates each piece of document information with an identifier for identifying the document information, and holds the document information. Therefore, the document generation unit 121 holds document information to be integrated and document information to be integrated. As shown in FIG. 6, document information to be integrated includes document information whose attribute values are characters and document information whose attribute values are numerical values. As shown in FIG. 7, the document information to be merged includes document information whose attribute values are characters and document information whose attribute values are numerical values.

After the process of step S102 is completed, the document generation unit 121 classifies the document information (step S103). Specifically, the document generator 121 classifies two pieces of document information based on regular expressions. For example, as shown in FIG. 8, when the document generation unit 121 holds the document information to be integrated and the document information to be integrated, the document generation unit 121 refers to the document information and determines the type of the attribute value to generate the two pieces of document information. Classify. The document generation unit 121 may classify the two pieces of document information according to whether the attribute value is numeric or not, or classify the two pieces of document information according to whether the attribute value is characters. As a result, as shown in FIG. 8, the document information of the character group including the document information to be integrated whose attribute values are standardized by characters and the document information of the integration destination, and the document information to be integrated whose attribute values are standardized by numerical values. It is classified into the document information of the numerical group including the information and the document information of the integration destination. The document generation unit 121 stores the document information of the character group and the document information of the numerical value group in the document storage unit 112 . As a result, the document storage unit 112 stores the document information of the character group and the document information of the numerical value group.

When the process of step S103 is completed, the cleansing unit 122 then determines whether the attribute value is a numerical value (step S104). More specifically, cleansing unit 122 accesses document storage unit 112 and acquires one piece of document information stored in document storage unit 112 . When acquiring one piece of document information, the document generation unit 121 confirms the attribute value of the acquired document information and determines whether the attribute value is a numerical value. If the attribute value is a numerical value, the cleansing unit 122 has acquired the document information of the numerical group. If the attribute value is character, it means that the cleansing unit 122 has acquired the document information of the character group.

In the process of step S104, when the cleansing unit 122 determines that the attribute value is not a numerical value (step S104: NO), it determines that the attribute value is a character, and extracts each attribute value aggregated into the document information of the character group. (Step S105). When the process of step S105 is completed, the cleansing unit 122 performs data cleansing on at least one of the extracted attribute values (step S106). For example, as shown in FIG. 9, the cleansing unit 122 performs data cleansing by re-expressing brackets expressed in half-width characters in full-width characters. The cleansing unit 122 also performs data cleansing to convert characters that depend on the environment of the computer into characters that do not depend on the environment. Furthermore, the cleansing unit 122 performs data cleansing to convert uppercase alphabets to lowercase alphabets. Note that these processes are an example of data cleansing, and may be changed, deleted, or added according to design.

When the process of step S106 is completed, the analysis unit 123 performs morphological analysis on the attribute values for which data cleansing has been performed (step S107). More specifically, the analysis unit 123 refers to the dictionary data stored in the dictionary storage unit 113 to perform morphological analysis. Dictionary data is data relating to a dictionary necessary for morphological analysis. By performing morphological analysis, some attribute values are separated into words as shown in FIG. That is, some attribute values are divided into words representing proper nouns and words representing common nouns, with a space between these two words. Depending on the attribute value such as secretary or manager, the attribute value itself may be a word.

When the process of step S107 is completed, the first extraction unit 124 extracts syntactic features (step S108). More specifically, the first extraction unit 124 extracts the syntactic feature quantity based on a known first extraction method for extracting the syntactic feature quantity together with the important words. A syntactic feature quantity represents a feature quantity relating to the syntax of a word. Therefore, the first extracting unit 124 extracts the first syntactic feature amount corresponding to the word from the document information to be integrated on which the morphological analysis has been performed. The first extraction unit 124 also extracts a second syntactic feature amount corresponding to a word from the document information of the integration destination on which the morphological analysis has been performed. Furthermore, the first extraction unit 124 extracts important words together with the extraction of the first syntactic feature amount and the second syntactic feature amount. As a result, as shown in FIG. 11, the first syntactic feature amount and the second syntactic feature amount corresponding to each word are extracted as numerical values, and important words are extracted. The first extraction unit 124 stores the extracted first syntactic feature amount, second syntactic feature amount, and important word in the syntactic feature storage unit 114 in association with each other. As a result, the syntactic feature storage unit 114 stores the first syntactic feature quantity, the second syntactic feature quantity, and the important word associated with each other.

Here, as the first extraction method described above, for example, Term Frequency - Inverse Document Frequency (tf-idf) (Ramos J. Using tf-idf to determine word relevance in document queries [C] // Proceedings of the first instructional conference on machine learning. 2003, 242: 133-142.) is known. tf-idf is the word frequency tf (Hans Peter Luhn, A Statistical Approach to Mechanized Encoding and Searching of Literary Information, IBM Journal of Research and Development. 1 (4): pages 309-317.) and inverse document frequency. idf (Karen Sparck Jones, A Statistical Interpretation of Term Specificity and Its Application in Retrieval, Journal of Documentation, Vol. 28 Issue: 1, pages 11-21 (1972)). The idf plays a role of lowering the importance of words appearing in many documents and raising the importance of words appearing only in specific documents. In the first embodiment, tf-idf is used as the first extraction method. Instead of tf-idf, for example, Best Matching 25 (BM25) (Robertson S, Zaragoza H. BM25 and beyond[J]. Foundations and Trends (registered trademark) in Information Retrieval, 2009, 3(4): 333-389.) may be adopted.

When the process of step S108 is completed, the second extracting unit 125 extracts semantic features (step S109). More specifically, the second extraction unit 125 extracts semantic feature quantities based on a known second extraction technique for extracting semantic feature quantities. A semantic feature quantity represents a feature quantity relating to the meaning of a word. A known second extraction method is stored in the vector storage unit 115 . A specific example of the second extraction method will be described later. The second extraction unit 125 can express a word by a vector corresponding to the word by using the second extraction method. Therefore, the second extraction unit 125 acquires the important words from the syntactic feature storage unit 114, and extracts, from the acquired important words, specific semantic feature quantities corresponding to the important words as vectors using the second extraction method. do. As a result, as shown in FIG. 12, semantic feature quantities corresponding to important words are extracted as vectors. The second extraction unit 125 associates the extracted semantic feature quantities with the important words and stores them in the semantic feature storage unit 116 . As a result, the semantic feature storage unit 116 stores the semantic features and important words associated with each other.

Here, for example, word2vec (Distributed Representations of Words and Phrases and their Compositionality Advances in Neural Information Processing Systems 26 (NIPS 2013)) is known as the above-described second extraction method. In the first embodiment, word2vec is used as the second extraction method. Vectors for Word Representation Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 1532-1543, October 25-29, 2014, Doha, Qatar.), etc. may be used.

When the process of step S109 is completed, the feature synthesizing unit 126 generates a synthesized feature amount (step S110). More specifically, as shown in FIG. 13, the feature synthesizing unit 126 weights the semantic feature stored in the semantic feature storing unit 116 with the first syntactic feature stored in the syntactic feature storing unit 114 to obtain a first synthesis. Generate features as vectors. Note that the first combined feature amount is the topmost combined feature amount among the three combined feature amounts shown in FIG. 13 . Further, the feature synthesizing unit 126 generates a second synthesized feature amount by weighting the semantic feature amount stored in the semantic feature storage unit 116 with the second syntactic feature amount stored in the syntactic feature storage unit 114 as a vector. The second combined feature amount is the combined feature amount located in the middle and the combined feature amount located at the bottom among the three combined feature amounts shown in FIG. 13 . In this way, by weighting the syntactic features to the semantic features, unmatching of syntactically inconsistent attribute names can be suppressed. After generating the first combined feature amount and the second combined feature amount, the feature combining unit 126 stores the generated first combined feature amount and second combined feature amount in the combined feature storage unit 117 . Thereby, the combined feature storage unit 117 stores the first combined feature amount and the second combined feature amount.

When the process of step S110 is completed, the similarity output unit 127 calculates the similarity (step S111). More specifically, the similarity output unit 127 acquires the first combined feature amount and the second combined feature amount from the combined feature storage unit 117, and calculates the cosine similarity between the acquired first combined feature amount and the second combined feature amount. do. Here, in the first embodiment, as shown in FIG. 13, the feature synthesizing unit 126 generates a second synthetic feature amount for the identifier "D2C1" and a second synthetic feature amount for the identifier "D2C4". Therefore, the similarity output unit 127 calculates the cosine similarity between the first combined feature amount related to the identifier "D1C1" and the second combined feature amount related to the identifier "D2C1". Further, the similarity output unit 127 calculates the cosine similarity between the first combined feature amount related to the identifier "D1C1" and the second combined feature amount related to the identifier "D2C4".

As a result, as shown in FIG. 14, the angle between the vectors based on the first combination of the identifiers "D1C1" and "D2C4" is the angle between the vectors based on the second combination of the identifiers "D1C1" and "D2C1". less than the angle of Therefore, the cosine similarity based on the first combination will be greater than the cosine similarity based on the second combination. That is, the former cosine similarity is closer to "1" than the latter cosine similarity. As a result, the similarity based on the first combination becomes relatively higher than the similarity based on the second combination, the attribute name "company" matches the attribute name "company name", and the attribute name "position" matches. It can be specified as not matching (ie, unmatched).

When the process of step S111 is completed, the similarity output unit 127 outputs the similarity (step S112). More specifically, the similarity output unit 127 outputs to the display unit 140 the matching list screen 10 including the calculated similarity, the attribute name to be integrated, and the attribute name to be integrated. As a result, the user viewing the matching list screen 10 can easily determine which attribute name should be integrated with the attribute name to be integrated. Details of the matching list screen 10 will be described later.

Next, a case where the attribute value is a numerical value in the process of step S104 described above will be described.

In the processing of step S104, when the cleansing unit 122 determines that the attribute value is a numerical value (step S104: YES), it subdivides the document information as shown in FIG. 15 (step S121). For example, based on regular expressions, the cleansing unit 122 subdivides the document information in the numerical group into document information in which attribute values representing dates are aggregated and document information in which attribute values representing dates and times are aggregated. Note that the cleansing unit 122 may skip the process of step S121 when it determines that there is no document information in which attribute values representing date and time are aggregated. In the first embodiment, since there is no document information in which attribute values representing dates and times are aggregated, detailed description of step S121 is omitted.

When the process of step S121 is completed, the cleansing unit 122 extracts attribute names (step S122). More specifically, as shown in FIG. 16, the cleansing unit 122 extracts attribute names associated with identifiers based on the first structured data D1, the second structured data D2, and the document information of the numeric groups. Since numeric attribute values are aggregated in the document information of the numeric group, the cleansing unit 122 compares these attribute values with the attribute values of the first structured data D1 and the second structured data D2 to obtain Attribute names of the first structured data D1 and the second structured data D2 can be extracted.

When the process of step S122 is completed, the analysis unit 123 performs morphological analysis on the attribute name extracted by the cleansing unit 122 (step S123). As in the process of step S107, the analysis unit 123 refers to the dictionary data stored in the dictionary storage unit 113 and executes morphological analysis. By performing morphological analysis, some attribute names are separated into words as shown in FIG. That is, some attribute names are split into two or more words with spaces between these two or more words. In the first embodiment, for example, "transfer date" is divided into two words "transfer" and "day". Depending on the attribute name such as "date of birth" or "birthday", the attribute name itself may be a word. Through the process of step S123, the analysis unit 123 acquires words included in the attribute name.

When the process of step S123 is completed, the second extracting unit 125 extracts semantic features (step S124). More specifically, the second extraction unit 125 extracts semantic features based on the above-described known second extraction technique. The second extraction unit 125 acquires a word from the analysis unit 123 and extracts a semantic feature amount corresponding to the word from the acquired word using the second extraction method. As a result, as shown in FIG. 18, the semantic feature quantity corresponding to the word is extracted as a vector. The second extraction unit 125 associates the semantic feature amount corresponding to the word with the word and stores it in the semantic feature storage unit 116 . Thereby, the semantic feature storage unit 116 stores semantic features and words associated with each other.

When the process of step S124 is completed, the feature synthesizing unit 126 generates an average feature amount (step S125). More specifically, as shown in FIG. 19, the feature synthesizing unit 126 divides the sum of the semantic feature amounts of one or more words included in the attribute name by the number of one or more words included in the attribute name. A first average feature amount and a second average feature amount are generated as vectors for each identifier. Note that the first average feature amount is the average feature amount located at the top among the four average feature amounts shown in FIG. 19 and the average feature amount located immediately below it. The second average feature amount is the lowest average feature amount among the four average feature amounts shown in FIG. 19 and the average feature amount located immediately above it. As described above, since the attribute values of the document information in the numeric group are numerical values, it is not effective to extract the semantic feature amount of the attribute values and generate the average feature amount. Therefore, the feature synthesizing unit 126 generates an average feature amount using semantic feature amounts of words included in attribute names. After generating the first average feature amount and the second average feature amount, the feature synthesis unit 126 stores the generated first average feature amount and second average feature amount in the synthesized feature storage unit 117 . Thereby, the combined feature storage unit 117 stores the first average feature amount and the second average feature amount.

When the process of step S125 is completed, the similarity output unit 127 executes the process of step S111 as shown in FIG. More specifically, the similarity output unit 127 acquires the first average feature amount and the second average feature amount from the combined feature storage unit 117, and calculates the cosine similarity of the acquired first average feature amount and the second average feature amount. do. Here, in the first embodiment, as shown in FIG. 19, the feature synthesizing unit 126 generates a first average feature amount related to the identifier "D1C2" and a first average feature amount related to the identifier "D1C3". Also, the feature synthesizing unit 126 generates a second average feature amount for the identifier "D2C2" and a second average feature amount for the identifier "D2C3".

Therefore, the similarity output unit 127 calculates the cosine similarity between the first average feature amount related to the identifier "D1C2" and the second average feature amount related to the identifier "D2C2". Further, the similarity output unit 127 calculates the cosine similarity between the first average feature amount related to the identifier "D1C2" and the second average feature amount related to the identifier "D2C3". As a result, as shown in FIG. 20, the angle between the vectors based on the third combination of the identifiers "D1C2" and "D2C3" is the angle between the vectors based on the fourth combination of the identifiers "D1C2" and "D2C2". less than the angle of Therefore, the cosine similarity based on the third combination is greater than the cosine similarity based on the fourth combination. That is, the former cosine similarity is closer to "1" than the latter cosine similarity. As a result, the degree of similarity based on the third combination is relatively higher than the degree of similarity based on the fourth combination, the attribute name "transfer date" matches the attribute name "salary date", and the attribute name "birth date" matches. date" (i.e. unmatched).

Similarly, the similarity output unit 127 calculates the cosine similarity between the first average feature amount related to the identifier "D1C3" and the second average feature amount related to the identifier "D2C2". Further, the similarity output unit 127 calculates the cosine similarity between the first average feature amount related to the identifier "D1C3" and the second average feature amount related to the identifier "D2C3". As a result, as shown in FIG. 21, the angle between the vectors based on the fifth combination of the identifiers "D1C3" and "D2C2" is the angle between the vectors based on the sixth combination of the identifiers "D1C3" and "D2C3". less than the angle of Therefore, the cosine similarity based on the fifth combination is greater than the cosine similarity based on the sixth combination. That is, the former cosine similarity is closer to "1" than the latter cosine similarity. As a result, the degree of similarity based on the fifth combination is relatively higher than the degree of similarity based on the sixth combination, the attribute name "date of birth" matches the attribute name "birthday", and the attribute name " Payday" can be specified as unmatched.

When the process of step S111 is completed, the similarity output unit 127 executes the process of step S112. That is, the similarity output unit 127 outputs the matching list screen 10 including the similarity and the like to the display unit 140 . As a result, the user viewing the matching list screen 10 can easily determine to which attribute name the attribute names to be integrated should be integrated, as in the case where the attribute values are characters.

FIG. 22 is an example of the matching list screen 10. FIG. The matching list screen 10 includes a screen related to the first structured data D1 and a screen related to the second structured data D2. Each screen included in the matching list screen 10 includes information to be integrated and information to be integrated. Specifically, each screen included in the matching list screen 10 includes identifiers, attribute names, and attribute values as information to be integrated. In addition, each screen included in the matching list screen 10 includes a corresponding identifier, a corresponding attribute name, and a corresponding attribute value as integration destination information. Each screen included in the matching list screen 10 includes candidate ranking and similarity between the information to be integrated and the information to be integrated. The candidate ranking represents the ranking of candidates for the integration destination with respect to the integration target. Similarity represents cosine similarity. A higher degree of similarity with a higher numerical value is assigned a higher rank as a candidate for integration destination. In this way, the matching list screen 10 is displayed on the display unit 140, and when the user confirms the matching list screen 10, the attribute name to be integrated can be easily determined with which attribute name.

As described above, in the first embodiment, the user confirms the matching list screen 10 and determines which attribute name should be integrated with the attribute name to be integrated. On the other hand, without displaying the matching list screen 10 on the display unit 140, for example, the similarity output unit 127 specifies the attribute name of the integration destination with the first candidate ranking, and the specified attribute name of the integration destination attribute names may be dynamically integrated by the similarity output unit 127 and the result may be displayed on the display unit 140 . This eliminates the need for confirmation work by the user, facilitating data integration.

FIG. 23 is a diagram briefly explaining the advantages of the first embodiment. The upper part shows the work of the user 20 according to the comparative example. The lower part shows the work of the user 20 according to the first embodiment. A user 20 such as a data scientist is required to analyze data and derive insights. Consumed in the processing process. In particular, in data integration in the preprocessing process, the expression of attribute names is often not unified, and each attribute name is individually checked to determine whether it is the same as the integration target for the integration target. . As a result, the remaining 20% of work time was required to analyze data and derive insights, making efficient data integration difficult.

However, according to the first embodiment, as shown in the lower part, if the user 20 checks the matching list screen (not shown in FIG. 23) displayed on the terminal device 100, it is possible to easily match and unmatch attribute names. can decide. As a result, it is possible to reduce the work time consumed in the preprocessing step and increase the work time required for deriving the insight.

As described above, according to the first embodiment, the terminal device 100 includes the storage unit 110 and the processing unit 120 . Storage unit 110 includes data storage unit 111 . The data storage unit 111 stores the first structured data D1 and the second structured data D2. The first structured data D1 includes attribute names and attribute values corresponding to the attribute names. The second structured data D2 includes a plurality of attribute names and attribute values corresponding to each of the plurality of attribute names.

On the other hand, the processing unit 120 includes a document generation unit 121 , a first extraction unit 124 , a second extraction unit 125 , a feature synthesizing unit 126 and a similarity output unit 127 . The document generation unit 121 generates document information that summarizes the attribute values included in the first structured data D1. Also, the document generation unit 121 generates document information in which a plurality of attribute values included in the second structured data D2 are aggregated. The first extraction unit 124 extracts the first syntactic feature amount from the former document information, extracts the second syntactic feature amount from the latter document information, and extracts the first syntactic feature amount from the latter document information, based on the known first extraction method described above. Specific key words are extracted in conjunction with the extraction of the syntactic features and the second syntactic features. The second extraction unit 125 extracts a specific semantic feature amount corresponding to the specific important word from the specific important word based on the known second extraction method described above. The feature synthesizing unit 126 generates a first synthesized feature amount by weighting the first syntactic feature amount to the specific semantic feature amount and a second synthesized feature amount by weighting the second syntactic feature amount to the specific semantic feature amount. do. The similarity output unit 127 outputs a first similarity indicating the degree of similarity between the first synthesized feature quantity and the second synthesized feature quantity, and a specific attribute name indicating an integration target attribute name determined according to the first similarity. output at least one of

This can assist efficient data integration. In particular, according to the first embodiment, not only the syntactic features of attribute names but also the semantic features of attribute values are considered. Don't miss attribute names that semantically match. On the other hand, since the processing related to data integration is not performed by focusing only on the semantic features of the attribute values, the amount of calculation can be reduced compared to the case where only the semantic features of the attribute values are noted. can.

(Second embodiment)
Next, a second embodiment of the present case will be described with reference to FIGS. 24 and 25. FIG. FIG. 24 is a diagram for explaining a matching example according to the comparative example. FIG. 25 is a diagram for explaining a matching example according to the second embodiment. In the second embodiment, instead of the first structured data D1 and the second structured data D2 described in the first embodiment, first semi-structured data D3 and second semi-structured data D4 will be used. .

Both the first semi-structured data D3 and the second semi-structured data D4 are electronic files whose data structures are defined in a tree structure format. Examples of such electronic files include eXtensible Markup Language (XML) files, JavaScript (registered trademark) Object Notation (JSON) files, and text files. That is, both the first semi-structured data D3 and the second semi-structured data D4 are data whose data structures are partially defined. In other words, both the first semi-structured data D3 and the second semi-structured data D4 partially have the structure definition of the data.

In the second embodiment, the second semi-structured data D4 corresponds to the integration target, and the first semi-structured data D3 corresponds to the integration destination. In the second embodiment, the data storage unit 111 stores such first semi-structured data D3 and second semi-structured data D4 as two structured data. Note that the first semi-structured data D3 and the second semi-structured data D4 only need to be stored in the data storage unit 111 before the process of step S102 described above is executed. The storage path of the second semi-structured data D4 is not particularly limited.

Here, as shown in FIG. 24, when integrating the attribute names included in the second semi-structured data D4 as the data set B into the attribute names included in the first semi-structured data D3 as the data set A, the second half A high-level attribute name (specifically, "chauffeur") included in the structured data D4 as an integration target and two high-level attribute names (specifically, chauffeur) included in the first semi-structured data D3 as an integration destination Specifically, matching with car information and driver information) is performed. At this time, the lower attribute name (specifically, series) included in the second semi-structured data D4 as an integration target and the four lower attribute names (specifically, series) included in the first semi-structured data D3 as integration destinations ("model", "price", "name", and "address").

However, even if the upper attribute name to be merged is syntactically matched with each of the two upper attribute names to be merged, the character forms do not match. It is difficult to accurately determine which of the higher-order attribute names matches. Similarly, even if the lower-level attribute name to be merged is syntactically matched with each of the four lower-level attribute names to be merged, the character forms do not match, so the lower-level attribute name to be merged is It is difficult to accurately determine which of the lower-level attribute names to be merged matches.

However, in the second embodiment, as shown in FIG. 25, the attribute value (specifically, X1, C63, etc.) corresponding to the lower attribute name included in the second semi-structured data D4 and the first semi-structured data Attribute values (specifically, Z2, M4, etc.) corresponding to lower attribute names included in D3 are used. In this way, if attribute values are used in the same way as in the first embodiment, both the syntactic features of attribute names and the semantic features of attribute values are taken into account, and lower-level attribute names to be integrated are integrated. It is possible to accurately determine which of the lower attribute names that are the first to match. In the second embodiment, as shown in FIG. 25, the lower attribute name "series" matches the lower attribute name "model" and does not match the lower attribute names "price", "name" and "address". (that is, unmatched).

In addition, when matching a higher-level attribute name to be integrated with a higher-level attribute name to be integrated, a lower-level attribute name and attribute value are used. In this way, if the lower attribute names and attribute values are used, all of the syntactic characteristics of the upper attribute names and lower attribute names and the semantic characteristics of the attribute values are considered, and the upper It is possible to accurately determine which of the attribute names of the higher level that is the integration destination matches the attribute name of . In the second embodiment, as shown in FIG. 25, the upper attribute name "chauffeur" matches the upper attribute name "driver information" and does not match the upper attribute name "car information" (i.e. unmatches). can judge.

As described above, according to the second embodiment, efficient data integration can be supported not only with the structured data described in the first embodiment, but also with semi-structured data. .

(Third embodiment)
Next, with reference to FIG. 26, a third embodiment of the present case will be described. FIG. 26 is an example of the data integration support system ST. The same reference numerals are assigned to the same components as those of the terminal device 100 shown in FIG. 3, and the description thereof will be omitted.

A data integration support system ST includes a terminal device 100 and a server device 200 . The terminal device 100 and the server device 200 are connected via a communication network NW. The communication network NW includes, for example, a Local Area Network (LAN) and the Internet.

The terminal device 100 includes an input section 130 , a display section 140 and a communication section 150 . On the other hand, the server device 200 has a storage section 110 , a processing section 120 and a communication section 160 . Both of the two communication units 150 and 160 can be implemented by the network I/F 100D or the short-range wireless communication circuit 100J. In this manner, the server device 200 may be used as a data integration support device by including the storage unit 110 and the processing unit 120 .

As shown in FIG. 26 , the storage unit 110 and the processing unit 120 described in the first embodiment may be provided in the server device 200 instead of the terminal device 100 . In this case, the input unit 130 of the terminal device 100 is operated, and the first structured data D1 and the second structured data D2 described above are sent to the storage unit 110 (more specifically, the data storage unit) via the two communication units 150 and 160. 111). First semi-structured data D3 and second semi-structured data D4 may be stored in storage unit 110 instead of first structured data D1 and second structured data D2.

The processing unit 120 accesses the storage unit 110, acquires the first structured data D1 and the second structured data D2, and stores the first structured data D1 and the second structured data D2 in the first embodiment. Executes the various processes described. Then, the processing unit 120 outputs the processing result to the communication unit 160 , and the communication unit 160 transmits the processing result to the communication unit 150 . The processing results include, for example, screen information capable of displaying the matching list screen 10 . Upon receiving the processing result, communication unit 150 outputs screen information to display unit 140 . Accordingly, the display unit 140 displays the matching list screen 10. FIG.

In this manner, the server device 200 may include the storage unit 110 and the processing unit 120 without the terminal device 100 including the storage unit 110 and the processing unit 120 . Further, the server device 200 may include the storage unit 110 and another server device (not shown) connected to the communication network NW may include the processing unit 120 . Even such an embodiment can support efficient data integration.

Although the preferred embodiments of the present invention have been described in detail above, the present invention is not limited to the specific embodiments, and various modifications can be made within the spirit and scope of the present invention described in the claims.・Changes are possible. For example, data cleansing does not need to be performed if the representation of attribute values to be subjected to data cleansing is unified in advance. Further, when the attribute value to be subjected to morphological analysis is written in words, the morphological analysis need not be performed.

Note that the following notes are further disclosed with respect to the above description.
(Appendix 1) First structured data including a first attribute name and a first attribute value corresponding to the first attribute name, a plurality of second attribute names, and the plurality of second attribute names a storage unit for storing second structured data including second attribute values corresponding to each; first document information in which the first attribute values are aggregated; and the second attribute values for each attribute name. Based on the first extraction method of generating aggregated second document information and extracting syntactic feature quantities representing feature quantities related to the syntax of words together with important words, the first document information is extracted from the first document information. extracting a first syntactic feature quantity of a word included in the attribute value of, extracting a second syntactic feature quantity of a word included in the second attribute value from the second document information, and extracting a second syntactic feature quantity of the word included in the attribute value of Based on a second extraction method of extracting a specific important word together with the extraction of the syntactic feature amount and the second syntactic feature amount, and extracting the semantic feature amount representing the feature amount related to the meaning of the word a first synthetic feature amount obtained by extracting a specific semantic feature amount corresponding to the specific important word from the specific important word, and weighting the first syntactic feature amount to the specific semantic feature amount; and to generate a second synthetic feature amount by weighting the second syntactic feature amount to the specific semantic feature amount, and representing the similarity between the first synthetic feature amount and the second synthetic feature amount a processing unit that outputs at least one of a first degree of similarity and a specific second attribute name indicating an integration destination of the first attribute name determined according to the first degree of similarity; A data integration support device provided.
(Appendix 2) After generating the first document information and the second document information, the processing unit, before extracting the first syntactic feature amount and the second syntactic feature amount, determining whether the first attribute value and the second attribute value are characters, and if determining that the first attribute value and the second attribute value are characters, the first syntax The data integration support device according to appendix 1, wherein the syntactic feature quantity and the second syntactic feature quantity are extracted.
(Appendix 3) The processing unit calculates the first similarity, and outputs the calculated first similarity to the display device together with the first attribute name and the second attribute name. The data integration support device according to appendix 1 or 2, characterized by:
(Appendix 4) After generating the first document information and the second document information, the processing unit, before extracting the first syntactic feature amount and the second syntactic feature amount, Determining whether the first attribute value and the second attribute value are numerical values, and when determining that the first attribute value and the second attribute value are numerical values, based on morphological analysis, A first word included in the first attribute name is obtained from the first attribute name, a second word included in the second attribute name is obtained from the second attribute name, and a second word is obtained from the second attribute name. based on the extraction method of extracting a first semantic feature quantity corresponding to the first word from the first word, and a second semantic feature quantity corresponding to the second word from the second word extracting a semantic feature amount, generating a first average feature amount by averaging the first semantic feature amount, and generating a second average feature amount by averaging the second semantic feature amount; a second degree of similarity representing a degree of similarity between the first average feature amount and the second average feature amount, and an integration destination of the first attribute name determined according to the second similarity; The data integration support device according to appendix 1, wherein at least one of the specified second attribute names is output.
(Appendix 5) The processing unit calculates the second similarity, and outputs the calculated second similarity to the display device together with the first attribute name and the second attribute name. The data integration support device according to appendix 4, characterized by:
(Appendix 6) Any one of Appendices 1 to 5, wherein both the first structured data and the second structured data include semi-structured data partially having a structure definition of the data. The data integration support device according to item 1.
(Appendix 7) First structured data including a first attribute name and a first attribute value corresponding to the first attribute name, a plurality of second attribute names, and the plurality of second attribute names Second structured data including second attribute values corresponding to each of them is stored, and first document information in which the first attribute values are aggregated and second document information in which the second attribute values are aggregated for each attribute name are stored. The first attribute value is extracted from the first document information based on the first extraction method of generating the document information of 2 and extracting the syntactic feature amount representing the feature amount related to the syntax of the word together with the important word. extracting a first syntactic feature of the word included in the second document information, extracting a second syntactic feature of the word included in the second attribute value from the second document information, and extracting the first syntactic feature of the word based on a second extraction method for extracting a specific important word in conjunction with the extraction of the amount and the second syntactic feature amount, and extracting the semantic feature amount representing the feature amount related to the meaning of the word, extracting a specific semantic feature amount corresponding to the specific important word from the important words of the above, and weighting the first syntactic feature amount to the specific semantic feature amount and the first synthetic feature amount and the specific A second synthetic feature is generated by weighting the second syntactic feature to the semantic feature of A data integration support method in which a computer executes a process of outputting at least one of a specific second attribute name indicating an integration destination of the first attribute name determined according to the similarity and the first similarity.
(Supplementary Note 8) The processing includes generating the first document information and the second document information and before extracting the first syntactic feature amount and the second syntactic feature amount. determining whether the first attribute value and the second attribute value are characters, and if determining that the first attribute value and the second attribute value are characters, the first syntactic The data integration support method according to appendix 7, wherein the feature amount and the second syntactic feature amount are extracted.
(Supplementary Note 9) The processing includes calculating the first degree of similarity and outputting the calculated first degree of similarity to a display device together with the first attribute name and the second attribute name. The data integration support method according to appendix 7 or 8, characterized by:
(Supplementary Note 10) The processing includes generating the first document information and the second document information and before extracting the first syntactic feature amount and the second syntactic feature amount. Determining whether the first attribute value and the second attribute value are numerical values, and if it is determined that the first attribute value and the second attribute value are numerical values, based on morphological analysis, the Acquiring a first word included in the first attribute name from the first attribute name, acquiring a second word included in the second attribute name from the second attribute name, and acquiring the second word from the second attribute name Based on the extraction technique, a first semantic feature quantity corresponding to the first word is extracted from the first word, and a second meaning corresponding to the second word is extracted from the second word extracting semantic features, generating a first average feature by averaging the first semantic features, and generating a second average feature by averaging the second semantic features a second similarity indicating the degree of similarity between the first average feature amount and the second average feature amount, and an integration destination of the first attribute name determined according to the second similarity The data integration support method according to appendix 7, wherein at least one of the specific second attribute names is output.
(Supplementary Note 11) The processing includes calculating the second degree of similarity and outputting the calculated second degree of similarity to a display device together with the first attribute name and the second attribute name. The data integration support method according to appendix 10, characterized by:
(Appendix 12) Any one of Appendices 7 to 11, wherein both the first structured data and the second structured data include semi-structured data partially having a structure definition of the data. The data integration support method according to item 1.
(Appendix 13) First structured data including a first attribute name and a first attribute value corresponding to the first attribute name, a plurality of second attribute names, and the plurality of second attribute names Second structured data including second attribute values corresponding to each of them is stored, and first document information in which the first attribute values are aggregated and second document information in which the second attribute values are aggregated for each attribute name are stored. The first attribute value is extracted from the first document information based on the first extraction method of generating the document information of 2 and extracting the syntactic feature amount representing the feature amount related to the syntax of the word together with the important word. extracting a first syntactic feature of the word included in the second document information, extracting a second syntactic feature of the word included in the second attribute value from the second document information, and extracting the first syntactic feature of the word based on a second extraction method for extracting a specific important word in conjunction with the extraction of the amount and the second syntactic feature amount, and extracting the semantic feature amount representing the feature amount related to the meaning of the word, extracting a specific semantic feature amount corresponding to the specific important word from the important words of the above, and weighting the first syntactic feature amount to the specific semantic feature amount and the first synthetic feature amount and the specific A second synthetic feature is generated by weighting the second syntactic feature to the semantic feature of Data integration support for causing a computer to execute a process of outputting at least one of a similarity and a specific second attribute name indicating an integration destination of the first attribute name determined according to the first similarity program.

100 terminal device 110 storage unit 111 data storage unit 120 processing unit 121 document generation unit 124 first extraction unit 125 second extraction unit 126 feature synthesis unit 127 similarity output unit

Claims (8)

first structured data including a first attribute name and a first attribute value corresponding to the first attribute name; a plurality of second attribute names; and a plurality of second attribute names corresponding to each of the plurality of second attribute names a storage unit that stores second structured data including a second attribute value;
generating first document information in which the first attribute values are aggregated and second document information in which the second attribute values are aggregated for each attribute name; extracting a first syntactic feature quantity of a word included in the first attribute value from the first document information based on a first extraction technique for extracting together with the important word, and obtaining the second document information; extracting a second syntactic feature of the word included in the second attribute value from, and extracting a specific important word along with the extraction of the first syntactic feature and the second syntactic feature and extracting a specific semantic feature quantity corresponding to the specific important word from the specific important word based on a second extraction method for extracting a semantic feature quantity representing a feature quantity relating to the meaning of the word. , a first composite feature obtained by weighting the first syntactic feature to the specific semantic feature, and a second composite obtained by weighting the second syntactic feature to the specific semantic feature generating a feature quantity, a first degree of similarity representing a degree of similarity between the first synthetic feature quantity and the second synthetic feature quantity, and the name of the first attribute determined according to the first similarity; a processing unit that executes a process of outputting at least one of a specific second attribute name indicating an integration destination;
A data integration support device comprising After generating the first document information and the second document information, and before extracting the first syntactic feature amount and the second syntactic feature amount, the processing unit generates the first document information and the second document information. determining whether the attribute value and the second attribute value are characters, and if determining that the first attribute value and the second attribute value are characters, the first syntactic feature amount and extracting the second syntactic feature;
The data integration support device according to claim 1, characterized by: The processing unit calculates the first degree of similarity, and outputs the calculated first degree of similarity to a display device together with the first attribute name and the second attribute name.
3. The data integration support device according to claim 1 or 2, characterized by: After generating the first document information and the second document information, and before extracting the first syntactic feature amount and the second syntactic feature amount, the processing unit generates the first document information and the second document information. determining whether the attribute value and the second attribute value are numerical values, and if it is determined that the first attribute value and the second attribute value are numerical values, based on morphological analysis, the first Acquiring a first word included in the first attribute name from the attribute name, acquiring a second word included in the second attribute name from the second attribute name, and performing the second extraction method a first semantic feature quantity corresponding to the first word is extracted from the first word and a second semantic feature quantity corresponding to the second word is extracted from the second word based on and generating a first average feature amount by averaging the first semantic feature amount, and generating a second average feature amount by averaging the second semantic feature amount, a second degree of similarity representing a degree of similarity between the first average feature quantity and the second average feature quantity; outputs at least one of the two attribute names,
The data integration support device according to claim 1, characterized by: The processing unit calculates the second similarity and outputs the calculated second similarity to a display device together with the first attribute name and the second attribute name.
5. The data integration support device according to claim 4, characterized by: both the first structured data and the second structured data comprise semi-structured data partially having a structure definition of the data;
6. The data integration support device according to any one of claims 1 to 5, characterized by: first structured data including a first attribute name and a first attribute value corresponding to the first attribute name; a plurality of second attribute names; and a plurality of second attribute names corresponding to each of the plurality of second attribute names store second structured data including a second attribute value;
generating first document information in which the first attribute values are aggregated and second document information in which the second attribute values are aggregated for each attribute name;
A first syntax of a word included in the first attribute value from the first document information based on a first extraction method for extracting a syntactic feature representing a feature relating to the syntax of the word together with important words. extracting a syntactic feature quantity, extracting a second syntactic feature quantity of a word included in the second attribute value from the second document information, extracting the first syntactic feature quantity and the second syntactic Extracting specific important words in conjunction with extracting features,
Extracting a specific semantic feature amount corresponding to the specific important word from the specific important word based on a second extraction method for extracting a semantic feature amount representing a feature amount related to the meaning of the word,
a first composite feature obtained by weighting the first syntactic feature to the specific semantic feature; and a second composite feature obtained by weighting the second syntactic feature to the specific semantic feature. produce the amount of
a first degree of similarity representing a degree of similarity between the first synthesized feature quantity and the second synthesized feature quantity; output at least one of the second attribute names;
A data integration support method in which a computer executes processing. first structured data including a first attribute name and a first attribute value corresponding to the first attribute name; a plurality of second attribute names; and a plurality of second attribute names corresponding to each of the plurality of second attribute names store second structured data including a second attribute value;
generating first document information in which the first attribute values are aggregated and second document information in which the second attribute values are aggregated for each attribute name;
A first syntax of a word included in the first attribute value from the first document information based on a first extraction method for extracting a syntactic feature representing a feature relating to the syntax of the word together with important words. extracting a syntactic feature quantity, extracting a second syntactic feature quantity of a word included in the second attribute value from the second document information, extracting the first syntactic feature quantity and the second syntactic Extracting specific important words in conjunction with extracting features,
Extracting a specific semantic feature amount corresponding to the specific important word from the specific important word based on a second extraction method for extracting a semantic feature amount representing a feature amount related to the meaning of the word,
a first composite feature obtained by weighting the first syntactic feature to the specific semantic feature; and a second composite feature obtained by weighting the second syntactic feature to the specific semantic feature. produce the amount of
a first degree of similarity representing a degree of similarity between the first synthesized feature quantity and the second synthesized feature quantity; output at least one of the second attribute names;
A data integration support program that causes a computer to execute processing.

Images