JP6710664B2 - Knowledge information management system, knowledge information management program, and knowledge information management method - Google Patents

Description

The present invention is a knowledge technology information management system capable of efficiently sharing knowledge technology necessary for performing those tasks and passing on know-how based on experience in product development/design/manufacturing departments and medical institutions of various manufacturing industries. And a knowledge technology information management program and a knowledge technology information management method.

In order for domestic companies to maintain the technological assets they have cultivated so far and connect them to the next generation, technology transfer is the most important issue that must be prioritized. Once technology is lost, it not only takes a considerable amount of time to acquire it again, but it is also feared that the quality problems that may occur during that time will seriously damage the creditworthiness of the company. In recent years, in particular, due to the large number of simultaneous retirements of generations that have supported Japan's growth, technology and knowledge may be lost rapidly.

For example, in design and development work, design intent is expressed in the form of drawings or specifications, but it is common for experienced engineers and young engineers to make a big difference in the contents. This is because there are differences in various aspects, such as the depth of "thinking" when considering design plans, the abundance of exhaustive extraction solutions for possible problems, and so on. As a method for filling these "differences in experience and knowledge", most companies will implement measures such as (a) business training and apprenticeship education, (b) manualization/standardization, and (c) technical document storage utilizing computer systems. It is carried out in. However, few companies replied that technology transfer was successful even after implementing these measures, and the effectiveness is not improving as expected. Therefore, there has been a strong demand for a mechanism that solves these problems related to technology transfer and allows anyone to easily obtain knowledge information.

Various proposals have been made to solve such problems. For example, in the following Patent Documents 1 and 2, a workflow in which the behavior of each worker is described in time series is used, and when the behavior of the worker is not included in the workflow after the content is changed, There has been proposed a management device and a management method capable of educating a non-skilled person about work by issuing an alarm to a person without direct guidance of a skilled person.

JP, 2012-146202, A JP, 2013-89122, A

By the way, work training and technology transfer in the apprenticeship system have been practiced for a long time, but nowadays the number of veteran engineers has begun to decrease rapidly, so there is no room for a long time, and one-on-one or small groups are targeted. Therefore, it is inefficient. In addition, it is difficult to repeat each time the need for a young engineer arises. In addition, even if it is manualized or standardized, information on why it is not described is not described, it is obsolete because continuous maintenance is not performed, it mentions that there is more know-how on the side of failure experience. There are many cases that have problems such as not being. With this, only the basic points of work can be conveyed, and sufficient technical transfer cannot be carried out.

On the other hand, there are many problems in the conventional knowledge information management utilizing a computer system. There are still many cases in which technical documents are digitized and centralized so that they can be collectively searched in one place, and document summaries and normalized keywords for improving searchability are not organized. Even if they are organized, it is necessary to read and check the document in the end, so it takes a lot of time and effort to reach useful information, and when busy, it gives up searching. .. The biggest problem is that no one knows about the existence of that information. It's the same as the fact that the precious technical assets are only dead.

Further, in recent years, attempts have been made to facilitate extraction of knowledge information by automatically extracting keywords for retrieval from internal and external information assets using AI technology, or organizing information into categories and creating a database. However, the AI technology itself is still in the developing stage, and an effective method of organizing information has not been provided, so sufficient results have not been obtained.

In addition, domestic manufacturing companies are accelerating their global expansion, and the problem of training local engineers to maintain the same product quality in Japan is being added to overseas bases. This takes much more work than in Japan, so there is a strong demand for a mechanism that will help to develop human resources overseas, overcoming the major obstacle of language differences.

Therefore, the present invention has been devised in order to solve these problems, and its main purpose is to store various tacit knowledge existing in a company as formal knowledge so that a veteran engineer can put it in mind. In addition to being able to reproduce the chained extraction process of knowledge related to the phenomenon, we can understand the acquired knowledge and design points in a short time and use it in our work. The present invention provides a new knowledge technology information management system, knowledge technology information management program, and knowledge technology information management method capable of easily sharing knowledge.

In order to achieve the above object, the present invention adopts the following configuration for managing knowledge information. That is, first, the knowledge information to be managed is accumulated by using the business term (word or idiom) as a pointer. Business terms that serve as pointers have controlled words, synonyms of the controlled words, and parallel translations of other languages to eliminate information redundancy, improve search hit rate, and enable information sharing in a global environment. To do.

Next, paying attention to the meaning of this business term, the business term category is defined, and a tree-like vocabulary structure from the highest-level concept to the business term is formed. Furthermore, the attribute information category necessary for managing the knowledge information associated with the meaning of the business term belonging to this business term category without fail is defined, and the basic data model which is the framework of knowledge information management is determined. This business term is collected from, for example, a regulation book or technical documents accumulated in a company, and corresponding attribute information is added/stored according to the attribute information category linked to the business term category to which it belongs. Then, by selecting this business term and performing a recursive search of the attribute information, the associated knowledge can be comprehensively acquired. More specifically, it has the following configuration.

In order to solve the above-mentioned problems, a first invention is a business term acquisition means for acquiring a business term, a control term for the business term for each business term acquired by the business term acquisition means, and a synonym for the control term. Business term registration means for registering at least one of a word and a parallel word, and a basic data model generation means for generating a basic data model centered on the business term for each business term registered by the business term registration means And a business term search means for searching any of a controlled term, a synonym, and a parallel translation of the business term acquired by the business term registration means, and outputting the search result together with attribute information according to the basic data model. The basic data model generation unit includes a vocabulary structure generation unit that generates, for each business term registered by the business term registration unit, a vocabulary structure defining at least a superordinate word and a subordinate word centering on the business term. And the attribute information linking unit that links the attribute information to the business term of each vocabulary structure generated by the vocabulary structure generating unit.

In a second aspect based on the first aspect, the business term search unit includes a recursive search unit that searches for a business term from the attribute information linked by the attribute information linking unit of the basic data model generation unit. It is a knowledge information management system characterized by that.

In a third aspect based on the first or second aspect, the basic data model generation means further includes an evaluation value assigning section that assigns an evaluation value to the attribute information linked by the attribute information linking section. Is a knowledge information management system.

A fourth invention is a business term acquisition step of acquiring a business term, and for each business term acquired in the business term acquisition step, a controlled word for the business term, or a synonym and a parallel translation word for the controlled term. A business term registration step of registering one or more; a basic data model generation step of generating a basic data model centering on the business term for each business term registered in the business term registration step; and the business term registration step A control term, a synonym, or a parallel word of the business term acquired in step S1, and a business term search step of outputting the search result together with the attribute information in accordance with the basic data model. The step includes a vocabulary structure generation step for generating, for each business term registered in the business term registration step, a vocabulary structure defining at least a superordinate word and a subordinate word centering on the business term, and the vocabulary structure generation step. The knowledge information management method is characterized by including an attribute information linking step of linking attribute information to the business term of each vocabulary structure generated in step.

According to a fifth aspect of the present invention, a computer uses a business term acquisition unit for acquiring a business term, a controlled word for the business term for each business term acquired by the business term acquisition unit, and synonyms and parallel translations for the controlled term. Business term registration means for registering any one or more of the above, a basic data model generation means for generating a basic data model centering on the business term for each business term registered by the business term registration means, and the business term The controlled term, the synonym, and the parallel word of the business term acquired by the term registration means are searched, and the search result is output together with the attribute information according to the basic data model. A vocabulary structure generation unit for generating, for each business term registered by the business term registration means, a vocabulary structure defining at least a superordinate word and a subordinate word for each business term registered by the business term registration means; The knowledge information management program is characterized by further causing the business term of each vocabulary structure generated by the vocabulary structure generation unit to further function as an attribute information connection unit that connects attribute information.

Terms that are attribute keys under the attribute information category are also managed as one of the business terms. This has an important meaning in the acquisition of knowledge information. Take the disc rotor of the disc brake as an example. Since the disk rotor has the function, principle, and method of "braking, motion stop, and friction method", "friction method" is selected and registered as an attribute indicating the function. On the other hand, the disk rotor has secondary functions of "heat generation" and "wear". These may be defined as the attributes of the disc rotor, but they are not inherently unique to the disc rotor, and are secondary to the "friction type" braking function. Therefore, by associating the knowledge of these sub-functions with the attribute of the function, it becomes common knowledge for parts having the same function.

When acquiring the knowledge information, it is possible to acquire the attribute of the subsidiary function of the relevant business term by recursively searching after acquiring the attribute of the disc rotor. In addition, the attribute information of the secondary function is automatically provided to the disc pad having a braking function, which is paired with the disc rotor, and the brake drum having the braking function in the brake unit of another mechanism. In this way, the managed business terms and the attribute information associated with them form a net-like knowledge structure by recursive search, and a large amount of knowledge information can be acquired in a chain.

The present invention also provides an effective management method of attribute information given to business terms. For example, for materials, it is necessary to evaluate and select various material properties in order to meet the requirements of products, so we would like to make side-by-side comparisons by unifying the evaluation axes of material properties. As a result of exploring a method that enables this, if the attribute associated with the broader word/lower word and the whole word/partial word of the “business term” is (a) , Both attributes (items) are equivalent and have a relation of inheritance (= law of attribute inheritance), (b) if the lower word is a classification of the upper word, the management attribute item of the upper word is the target There are rules that include attributes of business terms (= law of attribute inclusion), and (c) attributes (items) of whole words include attribute items of partial words (= law of attribute inclusion). By discovering, that is, referring to the relationship between the target business term and its broader terms, or the whole word of the target word, it is possible to guide the items of the vocabulary attribute linked to the target business term, and the labor of attribute definition is minimized You can keep it to the limit. This will be described in detail later.

When the attribute items are linked and managed according to this rule, all the attribute items set in the subordinate business terms are naturally accumulated for the superordinate business terms, and this is normalized during search. It will be available as a keyword. By using this, it becomes possible to easily perform a recursive (reverse lookup) search by side-by-side comparison such as finding a material suitable for a certain condition.

According to the present invention, a vocabulary structure centering on business terms is generated, and attribute information is associated with each business term and managed. It becomes possible to systematically acquire necessary and sufficient knowledge by gathering and networking them. Further, since the business terms are words or idioms, it is possible to read through them in a short time even when busy. In addition, the number of vocabularies to be managed is 10,000 to tens of thousands by narrowing down to business terms, which is far smaller than that for whole natural language. It is possible and easy to secure the response speed. Furthermore, simply preparing a bilingual word for each business term allows you to translate it into a foreign language, making it easy to use at overseas bases.

1 is a functional block diagram showing an embodiment of a knowledge information management system 100 according to the present invention. 1 is a hardware configuration diagram showing an embodiment of a knowledge information management system 100 according to the present invention. It is a flowchart figure which shows the flow of a business term acquisition and registration process. It is a flowchart figure which shows the flow of a basic data model generation process. It is a block diagram which shows a basic data model. It is a table figure which shows the specific example of a business term category. It is a table figure which shows the example of the relationship between a business term category and an attribute information category. It is explanatory drawing which shows the example of the attribute information linked|related with each business term. It is a table figure which shows the evaluation value of attribute information. It is explanatory drawing which shows the example of the attribute information linked|related with an item. It is explanatory drawing which shows the example of the attribute information linked|related with a material. It is explanatory drawing which shows the example of the attribute information linked|related with a function. It is explanatory drawing which shows the example of the attribute information linked|related with the characteristic and design requirement. It is explanatory drawing which shows the example of the attribute information linked|related with stress and an external factor. It is explanatory drawing which shows the example of an attribute expression type. It is explanatory drawing which shows the example of description for every attribute type. It is explanatory drawing which shows the example of description for every attribute type. It is explanatory drawing which shows the example of description for every attribute type. It is explanatory drawing which shows the example of description for every attribute type. It is explanatory drawing which shows the example of description for every attribute type. It is explanatory drawing which shows the example of description for every attribute type. It is explanatory drawing which shows the rule of attribute inheritance/inclusion. It is explanatory drawing which shows the example of the information acquisition by the recursive search of an attribute. It is explanatory drawing which shows the example of the information acquisition by the recursive search of an attribute. It is explanatory drawing which shows the example of the search by side by side comparison. It is explanatory drawing which shows the example of the search result by side by side comparison. It is an explanatory view showing an example of utilization for generation of a functional decomposition tree. It is an explanatory view showing an example of utilization for generation of a subsidiary function decomposition tree. It is explanatory drawing which shows the utilization example to a design contradiction analysis. It is explanatory drawing which shows the utilization example to a design procedure analysis. It is explanatory drawing which shows the utilization example to a design change influence analysis.

Next, embodiments of the present invention will be described with reference to the accompanying drawings. FIG. 1 is a functional block diagram showing an embodiment of a knowledge information management system 100 according to the present invention, and FIG. 2 is a hardware configuration diagram thereof. As shown in the figure, the knowledge information management system 100 includes an input unit 10, an information processing/storage unit 20, and an output unit 30.

The input unit 10 is operated by an operator to input various information such as business terms and search keywords. For example, the input unit 10 includes a keyboard as shown in FIG. 2 and a general-purpose input device such as a scanner or a touch panel. There is. On the other hand, the output unit 30 outputs various information such as an input result and a search result, and is composed of a general-purpose output device such as a monitor and a printer as shown in FIG. 2, for example.

As shown in FIG. 1, the information processing/storage unit 20 has various functions such as a business term acquisition means 40, a business term registration means 50, a basic data model generation means 60, and a business term search means 70. Specifically, the hardware (CPU 20a, CLC 20b, RAM 20c, ROM 20d, data bus 20e, interface (I/F) 20f, storage device 20g, etc.) of a general-purpose computer system as shown in FIG. 2 and its storage device 20g, etc. Each function is to be exhibited by the dedicated computer program (software) stored in.

The business term acquisition means (business term acquisition step) 40 acquires business terms (words or idioms) that serve as pointers, and the information sources thereof are various depending on each field. For example, in the manufacturing industry, bills of materials (BOMs), item lists, annotations in drawings, design item information obtained through design standardization and modularization activities, trouble information management, and defect prevention. Trouble occurrence point information obtained through activities, materials handbook, design handbook, catalogs of material manufacturers, etc. are effective information sources. As a specific method of acquiring the business term, in addition to manual input using a keyboard, a scanner, etc., uploading a file (document) converted into electronic data on a storage medium or a network (Internet, company LAN) N or There are ways to download.

The business term registration means (business term registration step) 50, for each business term acquired by the business term acquisition means 40, one or more of a controlled word for the business term, a synonym and a parallel word for the controlled term. Provide the function to register. Here, a controlled word for a business term means a business term used in a certain company that can be communicated only within that company or its affiliated company, or when it refers to the same thing but has a different expression method. Is a general term that controls the business term. For example, when gears have the same meaning such as gears or gears, bolts have screws, and screws have different expression methods, for example, gears and bolts are control words.

FIG. 3 shows an example of the flow of processing from the acquisition of business terms to the registration by the business term acquisition means 40 and the business term registration means 50. First, in the first step S100, a file (electronic data) serving as an information source is uploaded to the information processing/storage unit 20 or downloaded from the network N. Next, words are extracted from the document of the file imported in step S102. Note that the word extraction from various documents uses software having a function of extracting text data from a document file. Next, in step S104, words are extracted from the extracted text data using software having a morphological analysis function, and a business term is selected and registered from the extracted word group. Then, in step S106, if there is a business term that is synonymous with the extracted business term, it is defined, and one of them is registered as a controlled word and the other is registered as a synonym. In addition, bilingual words are also registered if necessary.

The basic data model generation means (basic data model generation step) 60 provides a function of generating, for each business term registered by the business term registration means 50, a basic data model centered on the business term. Further, the basic data model generating means 60 is composed of a vocabulary structure generating unit 60A, an attribute information linking unit 60B, and an evaluation value assigning unit 60C as shown in FIG. As shown in steps S200 and S202 of FIG. 4, the vocabulary structure generation unit 60A defines, for each business term registered by the business term registration means 50, at least a superordinate word and a subordinate word centering on the business term. The attribute information associating unit 60B provides a function of associating the attribute information with the business term of each vocabulary structure generated by the vocabulary structure generating unit 60A. Then, the evaluation value assigning unit 60C provides a function of assigning an evaluation value to the attribute information linked by the attribute information linking unit.

FIG. 5 shows an example of this basic data model, which is composed of a vocabulary structure and attribute information. As shown in the figure, the vocabulary structure is represented as a tree-like structure centered on the business term A and extending from its highest-level concept to the business term A. This is almost the same structure as the ontology known as a structured concept of natural language, except that the relationship between whole words and partial words is not retained in principle. In the business term, the relationship between the whole word and the partial word is the product configuration and recipe information, which is generally managed separately as the product configuration information (BOM). Since this is a part that changes depending on the product, it is meaningless to retain them for the purpose of universal knowledge information management. Therefore, the whole word-partial word relationship is not maintained except in special cases.

The business term A has a plurality of hierarchical structures, and the most abstracted concept is defined at the highest level. This is called a business term category. For example, in the case of the manufacturing industry, this business term category is, as shown in FIG. 6, "item", "part", "material/substance", "function/principle/method", "characteristic/requirement", "specification/variation". "Source", "design parameter", "state change/failure mode", "stress/external factor", "manufacturing process/construction method", "operation/work", "design/production/measuring equipment/device", "material" Properties, position, phase, color", "mode (sound, etc.)". Here, the "item" is a business term indicating a part or unit as shown in the figure, and is, for example, an O-ring or a gear. Further, the “site” is a business term indicating a shape element of a part, and is, for example, a sliding portion or a fastening portion. Further, "material/substance" is a business term indicating a material classification, type and name, and is, for example, metal or SUS304.

Further, each business term A is associated with attribute information indicating its meaning in more detail. For example, as shown in FIG. 7, if the business term is “item”, functions, sub-functions (=weaknesses/problems that should not be manifested originally), requirements, characteristics, specifications/specifications, design inputs, It is accompanying attribute information such as design parameters, design outputs, function loss factors, parts, rubbed items, and related items (related words). Further, in the case of a material, it is information such as a physical property value, a characteristic (good weather resistance, high toughness, etc.) or composition of the material. The highest level of these pieces of attribute information is called an attribute information category. As a result of analyzing the business terms, it was found that there is a universal relationship between the business term category and the attribute information category associated with it, and the association patterns are the same. In the present invention, this is utilized, and as shown in FIG. 7, a business term category and an attribute information category to be associated with the category are set in advance and used as a framework for organizing knowledge.

In the example of the vocabulary structure shown in FIG. 5, when the business term A is “car”, the upper term is “ground transportation equipment”, the lower term is “electric vehicle”, and the synonyms are “passenger car” and bilingual translation. The word is “Car”. The business term A is associated with attribute information such as requirements, characteristics, specifications/specifications, design inputs, design parameters, design outputs, function loss factors, parts, rubbed items, related items, etc. Has become.

The business term search means (business term search step) 70 searches for the business term registered by the business term registration means 50 or any of its controlled words, synonyms, and parallel words, and the search result according to the basic data model. Provides a function to output with attribute information. Further, as shown in FIG. 1, the business term search means 70 includes a recursive search section 70A, and reverses business terms from the attribute information linked by the attribute information linking section of the basic data model generation means 60. Provide a function to search by pulling.

Next, a specific example of the knowledge information management system 100 according to the present invention having such a configuration will be described with reference to FIGS. 8 to 31. First, FIG. 8 shows an example of attribute information associated with a business term as the center of a basic data model (vocabulary structure). Each business term has various attribute information to convey its meaning in more detail. Among them, universal attribute information is managed in association with the business term. As shown in the figure, in the case of a disk rotor as a business term, in addition to its outline and figure (photograph), universal attributes such as "function", "requirements", "secondary function", "failure mode", etc. Along with the information, specific examples of the respective attribute information are managed. For example, in the case of "function", specific functions such as "braking-motion stop-friction method" and "heat radiation-heat radiation plate method" are associated and managed. In the case of "Requirements", "Heat dissipation", "Heat resistance", "Brake braking stability" and "Impact resistance" are selected, and in the case of "Secondary function", "Frictional heat" and "Frictional noise" ], “wear”, and other specific functions are associated and managed.

In addition, the respective evaluation values are associated with the respective attributes and managed. FIG. 9 shows an example of the evaluation value. As shown in the figure, for example, in the “brake-motion stop-friction method” of the attribute information “function”, the highest evaluation 5 (having a strong relationship) of the five-level evaluation is associated and managed as an evaluation value. It Further, with respect to “heat generation” and “wear” of the attribute information “secondary mechanism”, the highest evaluation 5 (almost inevitably occurring) of the same five-level evaluation is linked as an evaluation value, and “shudder”, Regarding “corrosion” and “frictional noise”, evaluation 3 (occurs frequently) which is an intermediate of the five-stage evaluation is linked and managed.

FIG. 10 shows a relationship between an example of attribute information associated with a business term “item” and a specific example when the “item” is a disk rotor. When a disk rotor is searched as a business term, The attribute information and specific examples thereof are displayed together. In particular, in the manufacturing industry, the attributes associated with this “item” are the most, and as shown in the figure, there are more than 10 types, so in principle only universal attribute information is managed, and materials such as materials are decided at the time of design. It is desirable not to include attributes.

FIG. 11 also shows attribute information associated with the business term “material” and a specific example thereof. For this “material”, its characteristics, specifications, and composition are universal. It becomes an attribute. In the example in the figure, when you search for "polypropylene" for commercial use related to "materials", the properties such as "heat resistance", "alkali resistance", and "workability" are evaluated by those evaluation values ("◎" and " ○) etc.). In addition, detailed information such as “physical/mechanical properties” and “specific gravity” can be obtained as specification values. Furthermore, if synonyms are defined, synonyms such as "PP" and "polypro" will be displayed, and the tree structure of the basic data model will also be displayed.

FIG. 12 shows attribute information associated with the business term “function” and a specific example thereof. For this “function”, subsidiary functions, characteristics, and design parameters are universal attributes. Become. In the example of the figure, when the "function" is the "braking-motion stop-friction method", the secondary functions are "1. friction heat", "2. friction sound", "3. vibration", “4. Shear stress” is defined, and an evaluation value suitable for them is a five-level evaluation of the occurrence frequency (1. rare occurrence, 2. relatively occurrence, 3. occasional occurrence, 4. repeated occurrence) (5. Almost inevitably occurs) and is managed by being linked.

FIG. 13 also shows attribute information associated with the business term “characteristic/requirement” and a specific example thereof. This "characteristic/requirement" is an ambiguous wording such as "○○" or "○○", but it can be clarified by giving an attribute to the specification value that quantitatively indicates it. .. In the example of the figure, when cold district specifications are selected as requirements, specific cold resistance temperatures (lower limit value, reference value, upper limit value) are associated and managed as the specification values.

FIG. 14 also shows attribute information associated with the business term “stress/external factor” and a specific example thereof. The business term “failure mode” is generated by the presence of stress, and thus can be thought of as “implication” for stress. Therefore, as a related item of “stress/external factor”, “failure mode” is associated and managed. In the example shown in the figure, if the “stress/external factor” is high-frequency noise, it indicates that the related items such as malfunction and generation of induced current are also managed in association with the degree.

FIG. 15 shows an example in which the expression types of attribute information associated with business terms are classified into six. The attribute information associated with the business term basically holds the specific content in the relation between the title and the value, but can be classified into several expression types depending on the managed information. FIG. 15 is currently assumed, and Incomplete type, 2. Evaluation value retention type, 3. Numerical value retention type, 4. Numeric range retention type, 5. Threshold retention type, 6. There are six character string retention type expression types.

1. As shown in FIG. 16, the incomplete type is a type that has only an attribute title and does not have a specific numerical value or evaluation value. For example, as shown in the figure, when the attribute information of the disc rotor is “rubbing item”, the item has a specific numerical value or evaluation value such as “1. cylinder body, 2. piston, 3. brake pad”. Absent. 2. As shown in FIG. 17, the evaluation value holding type is a type that holds some evaluation value for the attribute title. Then, for example, when “1. frictional heat”, “2. frictional noise”, “3. vibration”, and “4. shear stress” are defined as “secondary functions” as shown in FIG. Is held as an evaluation value.

2. As shown in FIG. 18, the numerical value holding type is a type that holds quantitative numerical values and units for the attribute title. For example, when the attribute information of the disk rotor is “specifications” and the attribute is a thickness deviation, the thickness deviation is held as a numerical value. 4. The numerical range holding type is a type that holds the numerical range and unit for the attribute title as shown in FIG. For example, when the attribute information of the disk rotor is "specifications" and its attribute is specific gravity, the range of specific gravity is held as a numerical range.

5. As shown in FIG. 20, the threshold retention type is a type that retains the threshold, the unit, and the relationship with the threshold for the attribute title. For example, when the attribute information of the disk rotor is “heat resistant temperature” and the attribute is heat resistant temperature, the threshold value and unit are held. 6. The character string holding type is a type that holds a value expressed by a character string for the attribute title as shown in FIG. For example, as shown in the figure, when the attribute information of the disc rotor is "related item", it is held in a specific character string such as "1. antilock control, 2. skid prevention control, 3. traction control". To do.

Further, in the process of reaching the present invention, the present inventor establishes a certain rule (law of inheritance/inclusion of business term attribute) for each attribute of upper word-lower word and whole word-partial word of business term. I found that. First, between the high-order word and the low-order word, (a) if the low-order word is the type of the high-order word, both attribute items are equivalent and have a relation of inheritance. (B) If the lower word is a classification of the upper word, the rule that the upper word attribute item includes the lower word attribute item and the lower word inherits a part of the upper word is established.

For example, as shown in FIG. 22, if the upper word of the business term “metal packing” (lower word) is “rotary motion seal” and the lower word of “rubber seal” (upper word) is “oil seal”, Since "oil seal" is a type of "rubber seal", both attribute items (seal, oil resistance, chemical resistance, cold resistance, heat resistance, wear resistance, weather resistance, water resistance) are inherited. .. On the other hand, since “metal packing” is a category of “seal for rotary motion”, the attribute item of “seal for rotary motion” includes the attribute items of “metal packing” (heat resistance, pressure resistance). The "metal packing" inherits some of the attributes of the "rotary motion seal".

Next, regarding the whole word-partial word relationship, the following rules (a) to (g) are established, for example, when explaining the relationship between a unit (whole word) and a part (partial word). That is, the characteristics of the unit (a) are determined by the characteristics of the component group that constitutes the unit. The strength of the characteristic appears as the sum of the "characteristics" of each component. If there is a component that cancels the characteristic, the appearance of the characteristic is weakened or damaged. (B) In order to meet the "requirements" for a unit, there must always be a part in the configuration that has characteristics that match that requirement. Conversely, if there are no parts that have characteristics that meet the required specifications, the requirements will not be satisfied.

(C) The function of the unit is expressed by the function of the constituent parts. Also, the strength of the connection between the unit and the component affects the function to be manifested. (D) Similarly, the subsidiary function of the unit is expressed by the subsidiary function of the constituent parts. The strength of the connection between the unit and the component affects the function to be manifested. (E) When a function is defined for a unit, the part having that function becomes the main part of the unit. (F) Most of the characteristics of the unit are determined by the characteristics of the main parts that make up the unit. Conversely, if the characteristics of the main parts are insufficient, the characteristics of the unit are likely to be impaired. (G) The suppression of the secondary function to the unit is determined by the existence of the component having the function of suppressing the secondary function and the strength of the function. Conversely, when there is no component for suppressing the secondary function or the function is weak, the possibility of invention of the secondary function increases.

In the knowledge information management system 100 according to the present invention, by using such a data structure, not only a normal search but also information acquisition by recursive search of attributes can be easily performed. For example, as shown in FIG. 23, by recursively retrieving the attributes associated with the attribute information of the “item”, “sub-function” from “function” and “sub-function” from “requirements” “Specification value” and the like can obtain useful information such as “requirements/function/function loss affecting factor/specification” from “part”.

In the knowledge information management system 100 according to the present invention, a network of related business terms can be formed, and thereby more information can be acquired. For example, as shown in FIG. 23, when chloroprene is used as the rubber material, the material properties are excellent in “oil resistance” and “weather resistance”, and at the same time have predetermined specifications (cold resistance temperature, heat resistance temperature). There is. Therefore, if this is selected as the material for the O-ring at the time of design, "oil resistance" and "weather resistance" are reflected in the requirements, and the "cold resistance" and "heat resistant temperature" of the specifications are reflected as they are. Therefore, a meaningful design is possible.

In the knowledge information management system 100 according to the present invention, reverse lookup by side-by-side comparison is also possible. That is, according to the above-mentioned rule of inheritance/inclusion of business term attributes, it is possible to carry out reverse lookup by side-by-side comparison by propagating from upper word to lower word. FIG. 25 and FIG. 26 show examples of reverse lookup search by side-by-side comparison. First, as shown in FIG. 25, "plastic" is entered in the search window, and as a search condition, the check boxes for rigidity and cold resistance are checked, and the check boxes for evaluation values indicate those characteristics. Check "5. Very good" and "4. Good" and click the search button. Then, as shown in FIG. 26, “AES resin” is searched as one of the plastics satisfying the condition, and its control word, relational structure, synonyms, etc. are displayed in a table form as well as its characteristics and evaluation values. can do.

FIG. 27 shows another example of utilization of the knowledge information management system 100 according to the present invention, in which product configuration information and knowledge information are connected to generate a functional decomposition tree (functional system diagram) from a tree based on the product configuration. It is a thing. In the example shown in the figure, the product configuration of the disc brake is replaced with the functional configuration, and the function of each component and its relation can be understood at a glance. Further, FIG. 28 is a diagram in which a subsidiary function decomposition tree is generated using the subsidiary function attribute, and by generating this decomposition tree, the weakness inherent in the product can be grasped at a glance.

FIG. 29 is an example in which the knowledge information management system 100 according to the present invention is utilized for analysis of design point inconsistencies. As shown in the figure, by inputting the required specifications (cold area specifications) for the product and comparing them with the various attributes in the product configuration, it is possible to understand design problems. The figure shows an example of extracting the inconsistent portion from the relationship between the required specifications and the selected material.

FIG. 30 is an example in which the knowledge information management system 100 according to the present invention is utilized for design procedure analysis. As shown in the figure, if design input and design output are managed as knowledge information for each item, it is possible to analyze the design procedure by matching each step with product configuration information, and the same effect as Gantt chart creation can be obtained. can get. In addition, by creating such a design procedure analysis diagram, it is possible to analyze which part has to be reviewed when changing a part having the same logic as the design procedure tree expansion and present the information. .. In the example of FIG. 31, the brake pad in step 2 is changed, and if this is changed, it is possible to grasp at a glance which part of which process is affected.

As described above, in the knowledge information management system 100 and the knowledge information management method according to the present invention, the management target is narrowed down to the business term to capture its characteristics, form the vocabulary structure, and the knowledge information associated with the word. Attribute information for managing is managed in association with each other. Furthermore, by making the term indicating the item of attribute information one of the business terms to be managed, it is possible to expand the knowledge information by recursive search. Also, by propagating the attribute information of each business term to the upper word-lower word according to the above-mentioned rule of attribute inheritance inclusion, reverse lookup by side-by-side comparison is also possible.

As a result, despite the fact that each one is formed by short words or idioms, it becomes possible to systematically acquire necessary and sufficient knowledge information (knowledge) by gathering and networking them. Attribute information linked to business terms serves as a guideline when registering knowledge information, and not only promotes homogenization of information registration contents, but also provides normalized keywords, side-by-side comparison and search conditions. It can also be used to evaluate the degree of matching. And most of the business terms are words or idioms, so you can read them in a short time even if you are busy. In addition, the number of vocabularies to be managed is 10,000 to tens of thousands by narrowing down to business terms, which is far smaller than that for whole natural language. It is possible and easy to secure the response speed. Furthermore, simply preparing a bilingual word for each business term allows you to translate it into a foreign language, making it easy to use at overseas bases.

In addition, the knowledge information managed by the knowledge information management system 100 according to the present invention is matched with the product configuration information and the recipe information defined at the time of product development to create the functional decomposition tree as described above ( 27), creation of subsidiary function decomposition tree (FIG. 28), design inconsistency analysis (FIG. 29), design procedure analysis (FIG. 30), design change impact analysis (FIG. 31), failure prediction, etc. It is possible and can provide higher value. Further, as long as a computer program for realizing these on software can be prepared, the knowledge information management system 100 according to the present invention can be economically and easily realized only by using a general-purpose computer system such as a PC to input data.

100... Knowledge information management system 10... Input unit 20... Storage unit 30... Output unit 40... Business term acquisition unit 50... Business term registration unit 60... Basic data model generation unit 60A... Vocabulary structure generation unit 60B... Attribute information association Part 60C... Evaluation value assigning part 70... Business term search means 70A... Recursive search part

Claims (5)

Business term acquisition means for acquiring business terms,
For each business term acquired by the business term acquisition means, a control term for the business term, and a business term registration means for registering one or more of synonyms and parallel translations for the control term,
For each business term registered by the business term registration means, a basic data model generation means for generating a basic data model centered on the business term,
A control term of the business term acquired by the business term registration means, a synonym, a parallel term, and a business term search means for outputting the search result together with the attribute information according to the basic data model,
The basic data model generation means,
A vocabulary structure generation unit that generates, for each business term registered by the business term registration means, a vocabulary structure that defines at least a superordinate word and a subordinate word centering on the business term;
A knowledge information management system, comprising: an attribute information linking unit that links attribute information to the business term of each vocabulary structure generated by the vocabulary structure generating unit. In the knowledge technology information management system according to claim 1,
The knowledge information management system, wherein the business term search unit includes a recursive search unit that searches for a business term from the attribute information linked by the attribute information linking unit of the basic data model generation unit. In the knowledge technology information management system according to claim 1 or 2,
The knowledge information management system, wherein the basic data model generation means further includes an evaluation value assigning unit that assigns an evaluation value to the attribute information linked by the attribute information linking unit. A knowledge information management method as an information processing method performed by a computer under the control of software,
A business term acquisition step for acquiring business terms,
For each business term acquired in the business term acquisition step, a business term registration step of registering a controlled word for the business term and any one or more of synonyms and parallel translations for the controlled term,
For each business term registered in the business term registration step, a basic data model generation step for generating a basic data model centered on the business term,
A controlled term of the business term acquired in the business term registration step, a synonym, a parallel term, and a business term search step of outputting the search result together with attribute information according to the basic data model,
The basic data model generation step,
A vocabulary structure generation step for generating, for each business term registered in the business term registration step, a vocabulary structure defining at least its superordinate and subordinate words centered on the business term;
A knowledge information management method comprising: an attribute information linking step of linking attribute information to the business term of each vocabulary structure generated in the vocabulary structure generating step. Computer,
For each business term acquired by the business term acquisition means and for the business term acquisition means, a controlled word for the business term and any one or more of synonyms and parallel translations for the controlled term are registered. Business term registration means,
For each business term registered by the business term registration means, a basic data model generation means for generating a basic data model centered on the business term,
While functioning as a business term search means for searching any of the controlled term, synonym, and parallel word of the business term acquired by the business term registration means, and outputting the search result together with the attribute information according to the basic data model. ,
The basic data model generation means,
A vocabulary structure generation unit that generates, for each business term registered by the business term registration means, a vocabulary structure that defines at least a superordinate word and a subordinate word centering on the business term;
A knowledge information management program, characterized in that it further functions as an attribute information linking unit for linking attribute information to the business term of each vocabulary structure generated by the vocabulary structure generating unit.

Images