WO2006132000A1

WO2006132000A1 - Random traceability layer-based average contrastive analysis method and x-r control chart function unit program and random traceability layer-based average contrastive cause analysis program

Info

Publication number: WO2006132000A1
Application number: PCT/JP2005/018990
Authority: WO
Inventors: Masao Tsuruoka
Original assignee: Masao Tsuruoka
Priority date: 2005-06-10
Filing date: 2005-10-11
Publication date: 2006-12-14
Also published as: JP4172042B2; JP2006344193A

Abstract

A program which can monitor by highlighting an effect contrast, for each element in an accuracy region, that has conventionally been failed to be discriminated due to bias or mutual action even if a within-standard component is used, and can analyze causes for chronic characteristics deficiency problems during production on job site. The records of a traceability mixed production or an equivalent production in which the probable traceabilities of production processes, components and materials that are all the product quality elements ranging from materials to completed products are uniformly and randomly mixed are collected factitiously, the records are, as shown in flows in Fig. 1, are classified into layers based on the traceability discrimination of one element, one axis is added to a conventional X-R control chart, product problem characteristics average trend and average trend of element characteristics such as components sizes are monitored on the same graph, the correlation between those characteristics averages are confirmed, element characteristics items are auto-piloted, and the candidate cause of mostly correlated element characteristics is automatically judged and reported as well as in voice.

Description

Description 謇 Random traceability stratified mean contrast analysis method and X-R control chart functional part program and random traceability stratified mean contrast cause analysis program TECHNICAL FIELD The present invention relates to information processing and statistical processing for cause analysis. Regarding the program. Background Art In the cause analysis, it is necessary to record the characteristics of the elements together with traceability (birth) records of production elements such as parts, materials and process parameters, and operating elements such as fuel and equipment parameters. The records must be consistent. This traceability and identification management is also a guideline of ISO-9000, and it goes without saying that this management range and management accuracy have an effect on quality control. Need feedback.

With the advancement of design and simulator confirmation using CAD and high-precision measurement tools, the element accuracy has improved, and the occurrence of defects in the process that significantly divides the required standards has been reduced. With the development of products, the functions of the components have become complicated and subtle due to the multi-functional performance of products, and even if it is an element within the standardized standard, it is not visible due to interaction, etc. The standardization of the fuel design described in 0004 of the detailed description of the invention in Patent Document 1, where there are not many chronic problems that may be caused by being judged to be not included in the management target. However, there is a need to improve the statistical accuracy as described in paragraphs 0015 and 0038 of the detailed description of the invention in Patent Document 2. However, as in the method of Patent Document 2, even if all element characteristics are statistically only within the range of the soil σ, there is no offset partitioning of the bias and interaction. Yes.

What (What) the defect has occurred and why it has not occurred, where (Where) the defect has occurred and where it has not occurred, (When) the defect has occurred and when it has not occurred, and how ( In this aspect, we collect actual measurements and factual information, search for distinction points and change points, and propose a cause of the difference between distinction and change. Estimate several points, confirm the distinction points, change points, and whether there is any inconsistency with the information on those aspects, and make the one with no contradiction as the most probable cause as the universal cause in prayer There is a non-patent document 1 ATS method in which a bible procedure is standardized as a table for education and training.

This method is a method of analyzing the cause by effectively utilizing the expertise that can be possessed, and requires the user's own technical judgment. Even if information is collected and analyzed, as mentioned above, it can help even if it cannot be observed due to the minute fluctuations in the standard or the interaction of biased element characteristics. Even if you use the application software for each function alone to analyze the product characteristics, vs. element characteristics on a one-to-one basis, it will take many thoughts and errors to reach the cause while producing on-site. It takes time and effort. It is one method to narrow down the factors (candidates) from the logic and make engineers capable of analysis using the L12 direct experiment design method, etc. resident in the process, but it can be applied to any kind of product or industry. Usable, expert engineers, even those who have a little technical skill and are not good at logical considerations.Easy to operate with standard performance monitoring of conventional product performance characteristics. Therefore, there is a need for program software that can analyze the causes of chronic defects and troubles.

Patent Document 1: Japanese Patent Application Laid-Open No. 2001-153988

Patent Document 2: JP-A-2005-38098

Non-patent document 1: US, Kebna Itrigo, 1977, ATS (Analytics, Trouble, Showing) method

Disclosure of the Invention Problems to be Solved by the InventionStatistics of multiple factor elements that are not visible even after statistical processing due to minute variations such as within-standard fluctuations in product quality elements of parts, materials, and processes It provides an analytical method that prevents bias and highlights one-factor influence contrasts.

There are many independent programs for creating graphs such as standard deviation, correlation calculation, and scatter plots, and using them, the cause analysis work has become a little easier, but now it has become complicated and the analysis operation time is enormous. Is needed. Therefore, in conjunction with the above-mentioned method of contrasting the influence of factors, these existing single functions are systematically linked to monitor the performance of product performance characteristics in daily production management and approach the control limit. We provide a statistical cause analysis program that makes it easy to grasp the changes in product characteristics and the causal element characteristics of a chronic failure characteristic problem while monitoring the progress of the analysis with simple operations. Means for solving the problem

The traceability and identification processes, which are the management items in the ISO 9000 guidelines, are recorded in the quality control of parts, materials, processes, and finished products over the entire quality process, and the mutual relationship between components In order to offset the effects of action and bias and improve statistical accuracy, mass production that randomly mixes possible traceability in the production of parts, materials, and processes, and products with the same uniform random element traceability from accumulated records The quality records of the ATS method of Non-Patent Document 1 are not included in the calculation results of the same artificial simulator as the premise of the ATS method in Non-Patent Document 1, and only traceable fact data and actual measurement records of the same aspect as the ATS method Provide the basic data for the cause analysis of this program.

The interaction of non-confirmed elements is fixed at the center of possible mass production fluctuations, the fluctuations and bias effects of non-confirmed elements are offset, the accuracy of stratified statistics is increased, and the effect of confirmation element traceability is controlled. In order to make (contrast) stand out, the product quality record composed of the random element traceability is input to the database, and the identification of the element / raceability of parts, materials, and processes is divided into groups (keywords). Obtain the group average of the product inspection and operating output evaluation characteristics of the analysis target, repeat the traceability stratification for each element, and set the characteristics average of the analysis target. Traceability with the greatest inter-loop variation 'Identify parts, materials, and processes that are the main factor elements corresponding to the group. After that, with the same traceability stratification as when the identified main factor element was identified, the group average of each element characteristic such as the component size, material composition ratio, process condition parameter, etc. constituting the element and the problem characteristic to be analyzed Confirm the correlation between group averages and identify the element characteristics of the cause candidates that show the most correlation. Then, a regression evaluation is performed between the specified element characteristic average and the analysis target characteristic average, and the results are reviewed to determine the allowable standard and offset amount of the specified element characteristic to correct the occurrence and variation of the analysis target characteristic defect. It is used for the calculation of users.

In order to allow the computer to perform the above processing, identify cause candidates that are main factor element characteristics from random traceable element records, automatically answer them, verify and confirm them, and perform regression evaluation to take countermeasures for the causes The following six functions are built into this program: database function, secondary database function, statistical comparison table function, histogram function, X-R control chart function, and scatter chart function.

First, record the product serial number, machine number, and operation date, including the product serial number and machine operating characteristic data, the record of lacerability identification attached to the product characteristics, the parts incorporated into each product and operating equipment, Traceability identification record of material, fuel and process elements, characteristic data record of product performance inspection and equipment operation output evaluation result, and component dimensions, material and fuel component ratio linked to embedded element traceability identification record Record of element characteristics such as process or operating conditions, parameters, etc.

Establish a database function section that organizes and stores four types of records. Second, a secondary database function is provided to temporarily store the stratified data by sorting the database records into eyes (keywords) for traceability identification at each analysis stage. Third, in order to observe the peculiarities of secondary database records sorted from the database, a statistical comparison table is used to monitor the statistical standard deviation, average, maximum and minimum values of both database and secondary database records. A functional part is provided.

Fourth, in order to confirm the cause match, both the database and the secondary database are displayed on the monitor with the standard deviation and the estimated defect rate of the records, and a histogram function is provided to represent the frequency distribution of the problem characteristics to be analyzed. . Fifth, on the X-R control chart, not only the time axis of year, month, date and time, but also the traceability to be confirmed is the horizontal axis of the transition, and other characteristics such as product performance characteristics and operational output characteristics that are management characteristics In addition to the average transition, the average transition of element characteristics such as part dimensions, material and fuel component ratios, process and equipment operation setting condition parameters, etc. are displayed on the same graph, and the conventional single characteristic monitoring function is achieved. At the same time, it is possible to monitor the synergistic relationship with the transition of element characteristics, and monitor the main factor element characteristic name and correlation index (coefficient) of the cause candidate that shows the most correlation between the stratified averages of element characteristics and analysis target characteristics. An X-R control chart function is provided to respond by display, message diagram, and voice. Sixth, a scatter diagram function unit is provided that performs a regression evaluation between the stratified averages of the analysis target characteristics and the main factor element characteristics confirmed by the XR control chart function unit and displays them on a monitor. Effect of the invention The current state of nuclear power generation for which Patent Document 1 has been applied for, and the power plant relationship of Patent Document 2, are not necessarily related to the power plant in any industrial site, but it may not be carried as planned. It is a reality and the premise is to identify the cause of any countermeasure. It is possible to analyze the cause and cause of the accuracy range that did not exist, and to improve the production quality with accuracy. Regardless of pharmaceutical analysis, agriculture or industry, the cause of the chronic failure problem can be taken.

In addition, it can be used for business development in new fields with accurate factor analysis and identification during the development of new products, advancement of risk prevention in advance by clarifying the operating limits of individual equipment, and equipment Can be improved in performance and size, and contribute to the reduction of wasteful consumption of resources.

Now that PCs have become supermarkets, the spread of them has naturally become a daily tool, but on one side, there are parts that are not fully utilized, and the conventional X-R that could easily be damaged even on paper Control charts have been useful for monitoring product performance characteristics and other characteristic transitions, but monitor charts in the X-R control chart function section are the component characteristics that make up products, such as component dimensions, material and fertilizer composition ratios, and processes. By handling conditional parameters on a single table, the transition of element characteristics can be monitored instead of on the same graph, and the element characteristics that are the cause and cause of abnormalities and changes in product performance characteristics can be easily identified. Became. Therefore, it has become possible to manage production processes more efficiently without dividing products and industrial fields.

Since the characteristic values of the constituent elements differ for each product and device, the random traceability stratified average contrast analysis method is used to control the operation of devices with computers such as satellites, aircraft, automobiles, and LCD televisions. When there is no problem, set the parameters of all the devices in a regular manner, make random changes to the devices themselves, record and accumulate each output characteristic change, and correlate and correlate with the effects of individual device characteristics on all device output characteristics. Read and obey the requirements as a matter of course, and let the device itself set the parameters of all devices without difficulty, so that the relationship that the designer does not notice is also obvious, according to the operation history of the device, It is possible to work if there is a next-seat factor or substitute device only in the event of a crisis. Resulted in a life-prolonging effect, also help to such as efficiency improvement of automatic adjustment function in a similar manner production for automatic set fresh machine. Brief Description of Drawings Figure 1 Random Traceability Stratified Average Control Analysis Method and Cause Analysis Program

FIG.

Figure 2: This is the database function part monitor screen.

Figure 3: Monitor screen of secondary database function section.

Fig. 4: Statistical comparison table function section monitor screen.

Figure 5: Histogram function monitor screen. ,

Figure 6: X-R control chart function section monitor screen.

Figure 7: Scatter plot function section monitor screen. BEST MODE FOR CARRYING OUT THE INVENTION The general public can understand the analysis process, and the method and programming for use on a personal computer can be realized by utilizing traceability and identification information. Example 1 Production date and sample number of product traceability identification (1) shown in Figure 2 and component B. CV (part B model number) and component C. CV of component traceability identification item (2) The data that can be broadly divided into four types, namely, the characteristic D and characteristic E of the product characteristic item (3) and the dimension item of the component B of the component element characteristic item (4) of the component, This is an example of a real record taken for cause analysis on a monitor.

Traceability and identification management, which are the management items of ISO—9000 guidelines, are managed throughout the quality process, and what is the serial number of products, operating equipment, parts, materials, and fuel components (What) Code, number, date of production of when made, production location, production factory, line name, production process number, machine number and type of machine (1) and (2) are records that can be traced to 'Raceability' items, and how (Product) inspection characteristics records and operational evaluation characteristics records are (3) records. dimensions and component ratio and the material, quantity, also recorded ^[_ ¾S elements characteristic recording elements of operating condition parameters adjustment processing and operating machinery or equipment inside the device (4) + - Ru

In the traceability identification process shown in (2) and (2) of Fig. 2, the discrimination reading in the traceability identification process is performed by reading the barcode of the product, the packaging of the parts and materials, the IC tag, the label attached to the product, etc. (1) The characteristics records in (3) and (4) linked to the traceability in (2) and (2) have created a system in which traceability breakdown records for each quality process stage can be easily obtained via network mail. As with the dimensions of only the part B of the characteristic item (4), the element characteristic record only needs to be obtained only for the necessary elements at the time of analysis. Computerization is not necessarily a precondition for each recording, but if it is arranged, the acquisition time can be reduced in today's widespread network.

Flow diagram 1, mixed traceability production, refers to the components in Fig. 2 where data is collected to investigate the mold effect among the differences in the components, materials, and production process parameters that are components of the product. This is to collect a complete characteristic record of a product in which the difference in traceability such as the model number used in the production of each part of the item (2) is identified at random.

Create a product assembly instruction table that randomly incorporates element traceability in the production history, select records of products that match the combination of element traceability from the accumulated data, and if there is a shortage, randomly follow the product assembly instruction table It is an additional production with traceability factors and product characteristic records are collected.

Flow chart of analysis of basic records for analysis Fig. 2 of the screen monitoring the database function (part) recording example shows the characteristic items (3) that were constantly monitored as the product defect rate expanded. In order to analyze which parts and which dimensions cause the fluctuations in the characteristic E, the production traceability records of the parts are obtained and the database function is used. It is a datasheet screen of the name in this program input to (Part). In this program, the statistical range is specified by arbitrarily coloring the column (cell) of the No. and product characteristic item name in this Datasheet table.

The dimension record in (4) is the function that the database function (part) has after identifying the component B that is the most variable influencing element.Figure 2 (2) shows the traceability identification items of the component B that is randomly incorporated. ) In part B. CV (model number), another dimensional measurement record that was recorded during inspection of part B was copied, distributed, and written manually, and entered according to the random identification. In that case, it takes an enormous amount of time, but with the inherent functions, input can be done in an instant. Actually, there are many other items in (1), (2), (3), and (4), but Figure 2 is thinned out and illustrated as an example. Can do. In addition, since there are many TXT types of records on various servers, the table conversion incorporation function (part) is located in front of the database function (part), although it is not in the explanation flow diagram 1.

The secondary database function (part) in the flow diagram 1 is used to input the product quality records consisting of random element traceability stored in the database function (part) as shown in Figure 3 and the input field for the item name to be sorted (5) Parts entered in Β · Sort by CV item 'Production mold number of part B entered in keyword input field (6) is recorded in layers (keywords) and sorted as shown in the list of records (7) It is a function (part) that temporarily stores records classified by sort layer according to the analysis stage and monitors them in the SortSheet screen called this program. Fig. 3 Table (7) is an example. The number of item columns is thinned out, and the number of rows below is cut halfway. Actually, you can scroll up, down, left, and right to see the whole picture.

The statistical comparison table function (part) in the flow chart 1 is a secondary database as shown in Figure 4.

(SortSheet) and the original database (Datasheet), both product characteristics, output characteristics and element characteristics distribution deviations are confirmed numerically, and each characteristic is observed to observe the specificity of the sorted secondary database data. This is a function that monitors the upper and lower limits of the standard and the numerical table of standard deviation, maximum, minimum, and average of all characteristics. It can be said that this is a functional part for preparing before creating a histogram, as described below. This monitor can be scrolled up and down to see the whole data.

The histogram function (part) in the flow diagram 1 confirms the peculiarity of the characteristic distribution of the secondary database that has been sorted in advance with arbitrary traceability, so that the product characteristics and output characteristics in the statistical comparison table in Figure 4 described in the previous section Select and double-click the property name field (cell) for either of the element characteristics and the element characteristics. Together with the frequency distribution of the same characteristics of the original database, the histogram between the maximum and minimum values of the analysis target characteristics of the database (Figure 5) 8), Histogram between specified values of database target characteristics (9), Histogram between maximum and minimum values of target characteristics of secondary database (10), Histogram between specified values of target characteristics of secondary database (10) 1 1) alongside the standard deviation, maximum, minimum, average of both, and the estimated failure rate of the analyte characteristics of the database (12). It represents the rate (13) is a function to allow the monitor. Figure 5 is a histogram of product problem characteristics E, which is the result of sorting production records of part B with model number 2 into a secondary database, and 2 of the types of part B from 1 to 8 It can be seen from the fact that the estimated defect rate (13) of the analysis target characteristic in the secondary database is zero, indicating that the number type does not produce defects. In this way, the purpose is to confirm the distribution of the characteristics of the production characteristics with some traceability of factors of the factors and causes in the original population, and the consistency of the results of identifying the factors and causes. The purpose of use is to determine the availability of each traceability such as part lot and model number. The X-R control chart function (part) in the flow diagram 1 is, first of all, a replacement for the conventional X-R control chart. As shown in Fig. 6, not only the time series but also the display column of the identification item name on the horizontal axis of the graph transition (part B. CV shown in Fig. 14 is attached to the graph transition horizontal axis label (15). Same as the sort layer of the model number, with the element traceability of parts, materials, and processes as the horizontal axis of the transition, all element traceability is automatically referenced and the average transition of the analysis target characteristics of the problem (18) Monitor the fluctuations in product characteristics E, and identify the elements of parts, materials, and processes that correspond to the element traceability identification items that appear in the identification item name display column (14) on the horizontal axis of the most fluctuating graph. Second, specify the reference range of the element characteristic item that corresponds to the specific maximum influence element at the position to be set at the bottom of the scrolled screen that is not visible in the screen shown in Fig. 6, and the component dimensions in that range. , Material component ratio, process condition parameters The element traceability item is automatically referenced on the horizontal axis of the transition to identify the same traceability as when the largest influential element was identified. Confirm the correlation with the average transition (18), identify the cause candidate name of the most correlated element characteristic item in the input field (19) of the correlation confirmation element characteristic name, and select the analysis target characteristic and its element characteristic. Correlation index between

(22) is displayed below the graph, and the search results report message diagram for possible causes

Analyze and report as (23). In addition, if traceability records can be managed throughout the quality process in accordance with the ISOr-9000 guidelines, this example program can be used for daily production management instead of the conventional X-R control chart. If you keep a record of product traceability, built-in element traceability, and product performance characteristics in the datasheet, you can remove a few devices, Rapid identification of materials and processes is possible with this X-R control chart function. In addition, for the regression evaluation to review the standard with accurate accuracy, such as the component dimensions of the causal element characteristics, the material component ratio, and the process condition parameter settings, the product records produced with the random traceability elements already described are used. If the accumulation is insufficient, production is performed according to the random assembly instruction table.

The scatter chart function (part) in the flow chart 1 monitors the correlation state between the two characteristics confirmed by the X—R control chart function (part) as shown in the scatter chart in FIG. 7, and the correlation index R square value

(24) The higher regression line formula (26) (27) of (25) is used for the revision calculation of the allowable specification and offset amount of the element characteristics of the cause candidates and factors.

Industrial Applicability At present, it is natural to use a computer to control each related device of multiple output characteristics of an instrument, but this random traceability stratified average control analysis method is added. In addition, a self-free setting type control of the setting parameters of the element device in the equipment in the requirement condition priority type is constructed. For all devices such as satellites, aircraft, and LCD TVs, output characteristics data such as all outputs, speed, and brightness, input quantities such as fuel and power, and internal device parameter settings added at the time The environmental analysis data such as the ambient temperature, brightness, etc. are placed on a single table in the computer and this comparison analysis method is used. Change the parameter settings of all the constituent devices at the same time in random and full range, and try self-test to measure the influence on the multiple output characteristics of the device at the same time, or select and calculate the data of the same random configuration from the accumulation The output characteristics at the point in time, each parameter setting, and the correlation and regression evaluation are performed. Requirements such as output and efficiency are given top priority within the limits of the resulting capacity In accordance with the above, self-control of the parameter settings of each in-flight multi-factor device reduces the burden on each in-flight device and works efficiently to obtain a life-prolonging effect. Use a substitute device that can work even a little, surpass the ground, and use the records during that time for repair services and new designs. There are also various automatic adjustment assembly machines. For example, in the case of an optical lens with a multi-axis alignment and an optical ball axis alignment and fixing assembly of the lens frame, the actual adjustment time starts from the average space coordinates of the adjustment completion results. There is a method that uses the rule of thumb until then to start moving in the direction that was the shortest on average, but because the size and shape of each lens and lens frame are different, automation of the folding angle with many strays There are not a few cases where the merit of this becomes poor. Therefore, using the random traceability layer average contrast analysis method, the measurement space coordinate point of each axis randomly combined so that the moving average of each axis coordinate becomes the center of the axis movable range of the automatic machine, and the shortest connecting the space point The route is determined in advance, the effect of optical characteristics is measured for each spatial point for each pair of lens and lens frame, and the adjustment range is narrowed down according to the regression evaluation of each axis coordinate. In order to prevent strays, perform efficient automatic adjustment work quickly and quickly.

Claims

The scope of the claims

1. The production and operation records in which the traceability differences of the components that are components of the product and the equipment to be operated are randomly and evenly collected are collected, and the group records are layered using the identification of the traceability of one element as a keyword. In order to perform cause analysis by means of a random traceability stratified average control analysis method in which each group average of the analysis target characteristic and the element characteristic of one element is calculated, and correlation and regression evaluation between the averages are performed.

Organize and store 4 types of records: traceability records of target production or operation, traceability records of the elements that compose them, measured records of analysis target characteristics, and measured records of element characteristics linked to element traceability. Database function part,

Secondary database function unit for temporarily storing records stratified by sorting database records by traceability identification,

Database and, recording the _secondary database sorted optionally for the specificity comparison of statistics for all characteristics of both statistical comparison table function unit,

Analyze both the database and the secondary database that stores records sorted from the database with traceability of the identified cause elements in order to verify the causes and factors together with the statistical comparison table function part Histogram function section for comparative monitoring of frequency distribution of problem characteristics,

Transition of the X-R control chart for managing product and event evaluation and equipment operation output characteristic transitions The horizontal axis can be replaced with the traceability items that you want to check without being limited to time, year, month, and date. Monitoring changes in the management characteristics depending on the level, and using the secondary database function section to check the group-wide fluctuations of the management characteristics grouped by element traceability to confirm all element traceability. Automatically traces the changes in the management characteristics, identifies the traceability elements that cause the most fluctuations in the management characteristics, and adds not only the average transition of the conventional management characteristics but also an additional axis. It is shown on the same graph along with the average trend so that the synergistic relationship between the two can be monitored and monitored. A function for confirming the correlation between the averages is added, and the correlation index (coefficient) is displayed at the bottom of the graph, and the same in the secondary database is sorted by the traceability of the element whose management characteristics are most fluctuating. Automatic reference patrol of all element characteristic items of the element, clarify the correlation between the problem characteristic and each element characteristic of the same element, identify the cause candidate that is the element characteristic showing the maximum correlation, and correlate the correlation Report with monitor display, message diagram and voice with index (coefficient),

X 1 D control chart function part,

Scatter chart function section that performs recovery evaluation between the characteristics confirmed in the X-R control chart function section for the user to calculate the allowable value and offset amount of the cause and factor element characteristics

These six functional units are provided, and the processing results are displayed on the monitor, and the factors that are the most correlated element characteristics that affect the maximum to the analysis target characteristics of the variation and fluctuation problems described above are the main causes. Automatic analysis,

Random Traceability Stratified Average Control Cause Analysis Program

2. X—R control for managing product quality and device operation output evaluation characteristic transitions The horizontal axis of the confirmation transitions in the charts is not limited to the chronological year, month, day, date, and traceability. To monitor the changes in management characteristics due to differences in element traceability,

In addition, a single axis is added in addition to the average transition of one conventional management characteristic, and it is displayed on the same graph together with the average transition of element characteristics so that the synergistic relationship between the two can be monitored.

In addition, a function to confirm the correlation between the averages of both characteristics is added, and the correlation index (coefficient) is displayed below the graph

X-R control chart function program that is a component of the random traceability layered average control cause analysis program that facilitates factor analysis

3. Evaluation target and biased / impact of element characteristics

Record the traceability of each element and each element characteristic so that the statistical impact assessment of one element characteristic on the characteristic is not hindered, and create a production and operation record where the differences in element traceability are mixed evenly. Collected,

The records are divided into stratified groups with the identification of traceability of one element as an eye (keyword), and an average is calculated for each group of characteristics to be analyzed.

Since element traceability within a group that does not correspond to a group is random and uniform, the average of the element characteristics of each element within each group is always the same in the middle of the possible range of variation. The average effect of the element characteristics is the same,

The difference in the influence of each group average of the characteristics to be analyzed is only the difference of the group average between the group of element characteristics of the element of the group,

As a result, it is possible to clarify the influence contrast (contrast) on the analysis target characteristics for each element characteristic, and simply use the group average value of the analysis target characteristics and element characteristics to categorize conventional random information. Analysis of accuracy range that was not possible with statistical methods.This is a component of the mean-contrast cause analysis program by random traceability stratification that enables statistical evaluation of correlation and regression between target characteristics and element characteristics.

Random Traceability Stratified Average Control Analysis Method