HOKKAIDO UNIVERSITY
`
`..
`
`···"·
`
`··"
`Title
`
`Design and Implementation for Scientific Article Data Base
`
`Author(s)
`
`Saito, Tatsuki
`
`Citation
`
`~ ttffi;JH::k~I~l:'¥~1:iJf3'i:~~E = Bulletin of the Faculty of
`Engineering, Hokkaido University, 151: 19-34
`
`Issue Date
`
`1990-07-30
`
`Doc URL
`
`http://hdl.handle.net/2115/42240
`
`Right
`
`Type
`
`bulletin (article)
`
`Additional
`Information
`
`Instructions for use
`
`Hokkaido University Collection of Scholarly and Academic Papers: HUSCAP
`
`001
`
`Facebook Inc. Ex. 1213
`
`

`

`~tifii:ilf*~'f::c~n;~TJJfx¥Rf!r
`~ 151 ~ (lfli!<: 2 :tf.)
`
`Bulletin of the Faculty of Engineering
`Hokkaido University. No. 151 (1990)
`
`Design and Implementation for Scientific Article Data Base
`
`Tatsuki SAITO*
`(Received March 31, 1990)
`
`Abstract
`
`Scientific article information data base system ANGEL was designed based on the data
`base management system ADABAS.
`It consists of an interactive input and retrieval system,
`intelligent input system and a system for the generation of a relation matrix. These systems
`were implemented by NATURAL of ADABAS. ANGEL is available for other scientific
`Initially, ANGEL was developed as an article data base relating
`article data base systems.
`to the sculptural surface generation theory of CAD (Computer Aided Design)/CAM (Com(cid:173)
`puter Aided Manufacturing).
`ANGEL is now open for public usage and is used at Hokkaido University Computer
`In this paper, the design concept of the system and input system of bibliographic
`Center.
`items are described, and several results obtained from the system are discussed. The input
`object of ANGEL is scientific articles written in English.
`
`1. Introduction
`
`Considering valid use of a scientific article data base, as well as use of bibliographic
`information, it is important to use relational information among articles effectually in order
`to reproduce higher intellectual information, and it is the motivation of this paper. There(cid:173)
`upon, the necessity of a system to comprehend a research trend was examined and it was
`investigated what information makes the essence of research clear. Attention was paid to
`the various relations among scientific articles. The relations are expressed by a labeled
`-graph, in which an article is represented as a point and a relation between articles is
`represented as a line. The aim of this paper is to develop the methodology for analyzing
`characteristics of the graph.
`A citation relation among articles is an important information of relations. Accord·
`ingly, it is inputted with information of cited place because it expresses the significance of
`citation.
`A keyword relation also is one of the necessary information to represent the com(cid:173)
`If there are many common keywords between two
`monness of the content or the theme.
`articles, then these articles are interpreted to be treated with a similar theme or the same
`
`* Department of Precision Engineering, Faculty of Engineering, Hokkaido University, Sapporo 060,
`Japan
`
`002
`
`Facebook Inc. Ex. 1213
`
`

`

`20
`
`Tatsuki SAITO
`
`theme. These information among articles are the objective relations of information that are
`not influenced by the person to examine.
`Relational graphs of articles are made out from these various information using this data
`base system, and some relational matrices are created from these graphs. The matrices are
`similarity matrices which have real values. After setting a mathematical criterion accord(cid:173)
`ing to a purpose of processing, they are unified into one similarity matrix finally in order to
`clarify the property of the relational structure by cluster analysis. Thus configuration of the
`In consequence, the system is designed to attach impor(cid:173)
`system is the target of this paper.
`tance to an effectual input of these information.
`The data base system ANGEL consists of keyboard subsystem, bibliographic item
`extraction subsystem, retrieval subsystem, display subsystem and storage subsystem in the
`entire composition. There are some shared modules which are entities of subsystems.
`These subsystems were implemented in NATURAL on ADA BAS.
`
`2. Constitution of the Data Base ANGEL
`
`The constitution of the data base ANGEL is shown in Fig. L Bibliographic information
`of scientific articles is put into the data base system by manual operation directly or by
`optical character reader automatically through recording media like a floppy disk device.
`The most outside frame (thick lined) means the data base management system ADABAS that
`is the relational model type for main frame computer. The chained lined frame means the
`data base system ANGEL which consists of four parts. The first part (upper-left) is the
`interactive input and retrieval system, and the second part (upper-right) is the intelligent
`input system. The third part (middle) is the data storage (f.ile). The fourth part (lower) is
`the generating system of relation matrix among scientific articles in the data base. The
`interactive input and retrieval system, and the intelligent input system are constructed
`respectively by four subsystems, that is, keyboard subsystem, display subsystem, retrieval
`subsystem and storage subsystem. The keyboard subsystem of the interactive input and
`retrieval system is for interactive manual input or interactive retrieval from VDT (visual
`display terminal).
`On the other hand, the bibliographic item extraction subsystem of the intelligent input
`system is for automatic input from a recorded file. The display subsystem is for displaying
`processed results in the interactive input and retrieval system or in the intelligent input
`system.
`In the intelligent input system, the processed results are displayed in order to verify
`them.
`The retrieval subsystem is to check whether processing data overlaps recorded data or
`not, and the subsystem of the interactive input and retrieval system enables to retrieve by
`author name, title, keyword et cetera. The storage subsystem is to store not only biblio(cid:173)
`graphic data but also the logging data of processing date, processing type (store, update,
`delete), user- ID, the last article number or the last article- ID to be accessed. An example
`of the logging data is shown in Fig. 2.
`The generating system of the relation is to set up the initial relation matrix from a
`
`003
`
`Facebook Inc. Ex. 1213
`
`

`

`Design and Implementation for Scientific Article Data Base
`
`21
`
`INTERACTIVE
`
`INPUT
`
`SYSTEM
`
`~etrieval Subsystem
`
`(Data
`
`--:-:-,
`-
`r - -
`~cquisitlon SubsystemJ
`~=--J=r==-=-
`~-.Display Subsystem J
`-===ci-=--=--=-,
`---J-r----...1
`r - - - -
`-----,
`~.!2{a~;~u~~~m_..J
`~----------~-----+--~ L--r------------------J
`Stora~o (---..1 GENERATING
`
`INTELLIGENT
`
`INPUT
`
`SYSTEM
`
`SYSTEM OF
`RELATION
`MATRIX
`
`'------·--t----1---ANGEL·-------l---l----J
`
`~-------------+----~DBMS/ADADA&----------t-----r----~
`
`Fig. 1 The System constitution of the data base ANGEL.
`
`specified relation and to output it into file.
`
`3. Scientific Article Information Input System of the Data Base
`
`3. 1 Acquisition of scientific article information
`There is a choice to purchase a marketing article data base to acquire bibliographic
`items by extracting a data field mechanically according to the specified format, and such
`choice is popular in general. There is also a choice to input bibliographic items manually
`after printed strings of article is read automatically by an optical character reader. A
`present situation of data input into data base needs manual inputting until when the possibil(cid:173)
`ity of the automatic item extraction realizes and the latter choice is available practically.
`The researcher data base needs as much as possible detailed information. Accordingly,
`sufficient information cannot be obtained only from the marketing bibliographic data base
`like the in former method, and then a system to input necessary information directly from an
`
`004
`
`Facebook Inc. Ex. 1213
`
`

`

`22
`
`Tatsuki SAITO
`
`Fig. 2 An example of the logging data in ANGEL.
`
`MORE
`PAGe
`
`245
`
`88·09·28
`
`I 6: I 2: 57
`PROCESS
`.HTMO
`YY!I\IDD
`ARTMOLST
`LASTUID
`TIMET
`-------- -------- -------- ----------- ----------- -------- -------- ----------
`CAD
`STORE
`3 0 I 2
`88·08·23 1I:53:54.6
`3 0 I 2 X10044
`CAD
`STORE
`3013
`88-08-23 11:54:35.7
`3 0 13 Xl0044
`CAD
`UPDATE
`·3014
`3013 X10044
`88-08-23 12:03:43.6
`CAD
`UPDATE
`3013
`88-08-23 12:05:02.8
`30 I3 X10044
`CAD
`STORE
`-3 0 14
`3 0 I 4 X!0044
`88-08-23 12:06:30.5
`CAD
`DELETE
`-3014
`3014 Xl0044
`88-08-23 12:08:23.7
`CAD
`UPDATE
`3014
`3013 XI0044
`88-08-23 12:12:58.0
`CAD
`UPDATE
`3014
`88-08-23 I2:I4:58.2
`3 0 I 3 Xl0044
`CAD
`STORE
`3014
`3 0 I 4 XI0044
`88-08-23 12:I5:20.5
`CAD
`STORE
`3015
`30 I 5 X!0044
`88-08-23 I2:15:36.7
`CAD
`STORE
`3016
`30 I 6 XI0044
`88-08-23 12:I6:03.3
`CAD
`STORE
`3017
`3017 X10044
`88-08-23 12:I6:20.6
`CAD
`STORE
`3018
`3018 X10044
`88-08-23 I2:I6:44.4
`STORE
`CAD
`3019 XI0044
`88-08-23 I2:17:35.6
`3019
`CAD
`STORE
`3020
`3020 Xl0044
`88-08-23 12:17:54.3
`CAD
`STORE
`3021
`88-08-23 12: 18: 14. 2
`3 0 2 I X10044
`CAD
`UPDATE
`3022
`3 0 2 1 X!0044
`88-08-23 12:48:35.8
`CAD
`STORE
`3022
`88-08-23 I2:48:50.2
`3022 X10044
`
`F :!A liB
`
`P.~SSWD
`
`original article is indispensable. A researcher cannot offord to the mounting many cost
`easily for inputting. Therefore, a researcher is required to input the data by himself, or the
`data is inputted by a part-time assistant. The system that enables to input data efficiently
`even if by nonprofessional key operator is necessary. For this reason, it is necessary to
`decrease the hitting frequency as well as possible at the time of data input in order to make
`key operation easy. Consequently, two systems using registered information in the data
`base were programed. The first system is interactive input and retrieval system, and the
`second is intelligent input system.
`
`3. 2 Interactive input and retrieval system
`When the input systems were designed, the following points were considered ;
`1) Improvement of input working efficiency.
`Especially, for a user who is not an professional key operator, to improve the efficiency
`of input work is effective.
`2) Reduction of misinput rate.
`In case of inputting abundent data, it is said generally that even professional key
`operators misinput at about 3 % rate.
`In consequence, the reduction of misinput is an
`important theme to acquire the information in order to minimmize errors. So data were
`processed on the basis of the next point.
`3) Valid use of registered data.
`To consider an input processing or a retrieval processing for the researcher database of
`scientific article, both a parent (citing) article and a child (cited) article are equally signifi(cid:173)
`cant, then it is necessary to treat them as independent articles because the database is to
`process various relations among articles. Accordingly, the same article appears many times
`as the reference which has a similar theme of research. Hence a tendency that the article
`by the same author appears frequently is remarkable in accordance with the increase of
`registered articles.
`
`005
`
`Facebook Inc. Ex. 1213
`
`

`

`Design and Implementation for Scientific Article Data Base
`
`23
`
`The necessity to check the duplication of inputted data that may be overlapping regis(cid:173)
`tered data comes into existence, and the duplication check in which the surname of the first
`author is the first retrieval key is valid in reliability and efficiency. Retrieval items are
`author name, title, article number, article identification number, time of publication, citing
`the authors name, journal name and key word. A retrieval condition can be set by also the
`logical AND or OR, that is, (surname of author) OR (title), (surname of author) AND (title or
`It can be used also intactly for
`other bibliographic information as occasional demands).
`simple retrieval processing. A full name or an affiliation of an author is stored or renewed
`only when these information appears as a parent article. On registering an article, it is
`always checked whether the article-ID overlaps the registered article-ID or the same title
`has been stored in the database. Consequently, the duplication of data can be prevented, and
`a consistency of data can be established simultaneously with the exclusion of the uselessness
`of a storage area.
`When necessary data is displayed in the retrieval screen, it can be shifted to the
`registration mode at once in the interactive input and retrieval system. Therefore, after
`updating, it is possible to register newly by using a part of the data that has been registered.
`It can be also updated when the affiliation of an author changes. Becoming aware of error
`occasionally after a screen proceeds while interactive data inputting, it is possible to send the
`screen backward in order to correct errors before it is stored in the database.
`
`3. 3 Details of interactive input and retrieval system
`An input menu form is adopted for ease of viewing and for restraining the input error.
`The information that has been inputted are preserved for reuse as much as possible. The
`MAP function of the screen editor of NATURAL was applied to screen design.
`It is a
`convenient function to make screen layout flexible. Fig. 3. 1 shows the screen of selecting
`ANGEL among the data base of HUCC. The screen of starting ANGEL is shown in Fig.3.
`2. A file name to be accessed, a password and selection parameter of processing are inputted
`here. A screen to input the selection whether a parent article or not is displayed after this
`If the user has the right of writing the data base. After 'Y'
`starting screen is displayed.
`-key is hit when the article is parental, or after only 'send'-key is hit when otherwise, the
`retrieval screen of Fig. 4 is displayed. An author name and a title are inputted in this screen
`as a retrieval item. Because the same article appears frequently in the item of a reference,
`(1)
`to avoid duplication of registration of the same data,
`(2)
`to check error of the registered data,
`retrieval processing is executed first.
`The default condition of retrieval is only a surname of an author. However the retrieval
`condition may adopt logical OR of each item or its logical AND. From the viewpoint of the
`efficiency, it may be advantageous not to input the title because it is long in comparison with
`an author name and it is inaccurate sometimes in case of being cited. To hit 'send'- key in
`this screen after necessary item is inputted, according to the specified condition, the retrieval
`If a corresponding article is registered already, then the data of
`process is executed.
`
`006
`
`Facebook Inc. Ex. 1213
`
`

`

`24
`
`Tatsuki SAITO
`
`( 90-02-07 )---------------------
`
`------------------------ AVAILABLE DATABASES
`F DATABASE
`CONTENTS
`INFORMATION SYSTEM
`AIRIS
`A. I. & I. R. DOCUMENT
`ALGO
`AN ALGORITHM
`INFORMATION SYSTEM
`ALTS
`AGRICULTURAL LONG-TERM STATISTICS DATABASE
`S ANGEL
`Article and Graphic Engineering Library
`COGBASE Cognitive Science Data Base
`FRM
`FERROELECTRICS AND RELATED MATERIALS
`HEAD
`HOKKAIDO UNIVERSITY ECONOMIC AND ACCOUNTING DATABASE
`HGEN
`GENETIC
`INFORMATION DATABASE
`HTCS
`Heat Transfer and Combustion Symposium database
`MEDRAD
`MEDICAL RECORD DATABASE OF RADIOLOGY
`NRDF
`CHARGED PARTICLE NUCLEAR REACTION DATA FILE
`QCLDB
`QUANTUM CHEMISTRY LITERATURE DATA BASE
`RRR
`RESOURCES FOR ROAD RESEARCHERS
`SESS
`SOVIET ECONOMIC STATISTICAL SERIES
`
`SELECT FUNCTION G :DISPLAY GUIDE-BOAD
`N :DISPLAY NOTES
`S :CALL THE DATABASE
`E :END
`
`Fig. 3-1
`
`fiELCOWE
`
`TO
`
`DATABASE
`
`##
`########
`#######
`##
`##
`##
`##
`##
`II#
`##
`###
`###
`II#
`##
`#
`II#
`## #
`##
`##
`#II
`#######
`##
`#
`##
`##
`##
`#
`##
`##
`#####
`##
`#########
`##
`# ##
`##
`##
`#It
`##
`##
`#
`##
`###
`#######
`########
`#######
`##
`##
`##
`#
`*******************************************************************************
`*
`(MPSU20Z] *
`FILE-NAME,
`PASSWORD AND
`PROCESS-SELECTION INPUT WENU
`*
`*
`*
`*
`*
`*
`*
`*
`*
`*
`*
`*
`*
`*
`*
`*
`*
`*
`* •
`*
`*
`*
`*
`*
`*******************************************************************************
`
`USER-ID
`
`; Xl0044
`
`CAD
`
`(+FILENAME)
`FILE-NAME
`(+PASSWD)
`PASSWORD
`(+SELPRO!)
`PROCESS
`, LIST==>L,
`INPUT TO PROCESS=•>
`PROCESS:
`INTERACTIVE RETRIEVAL ONLY==>! )
`FIN [ QUIT FROW ANGEL ] ==> PFI!
`
`Fig. 3-2 Start up screen of ANGEL.
`
`registered content is displayed as in Fig. 5.
`In case that there are plural data of the same author, to hit 'send'-key successively,
`corresponding bibliographic data are displayed in turn as in Fig. 6. Finding an error in the
`display screen, to hit 'U' at the last argument, it is possible to input the correct data instead
`of the error data.
`In case of registering a new article of the same author while retrieving,
`it is advisable hit 'N' at the last argument, then it is possible to reuse intactly the cor(cid:173)
`responding data that has been inputted in the database, and it is necessary for the keyboard
`to only put data of the field that has never inputted.
`While this method gives a few advantages at the initially creating stage of the data base,
`the effect can be expected in the practical period after numerous data are stored in the data
`base. Especially in the case of a researcher data base, it is not rare that the same author
`appears frequently in case of the treatment of comparably narrow field. Consequently, the
`
`007
`
`Facebook Inc. Ex. 1213
`
`

`

`Design and Implementation for Scientific Article Data Base
`
`25
`
`******************************************************************************
`*
`.
`(~PSR5-IO]•
`* RETRIEVAL MENDE OF' REGISTERED ARTICLES
`*
`FOR FILB
`: CAD
`*
`*
`******************************************************************************
`*
`*
`* •
`LAST LOGON-UID
`: X10044
`DATE
`:
`*
`LAST ARTNO = 3028
`*
`* -------------------------------------------------------------------------- *
`* ( RETRIEVAL CONDITION
`*
`: OR==>
`. AND==>A, CONTINUE==>C. QUIT==>Q )
`• -------------------------------------------------------------------------- *
`(SURtiAifE)
`*
`(FIRST, MIDDLE !!Allll)
`*
`AUTHOR:
`COONS
`*
`*
`*
`*
`*
`*
`*
`*
`*
`******************************************************************************
`
`TITLE:
`
`Fig. 4 Retrieval screen for author and title.
`
`++++++++++++++++++++++++++++++++++++++++++++++++++++t+++++++++t+++++t++t+++++++
`THE FOLLORING ARTICLE ALREADY REGISTERED II
`[~PDSSED]
`
`+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
`ARTICLE-NO
`ARTICLE-ID COONS(l974)
`==============~================================================================
`TITLE
`:
`SURFACE PATCHES AND B-SPLIIE CURVES
`
`AUTHOR 1
`AFFIL. NO.
`
`COOMS
`l
`: AFFILIATION
`
`S. A.
`,
`SYRACUSE Ul!IV.
`
`AFFIL. ~0.
`
`01
`
`JOURNAL COMPUTER AIDED GEO~ETRIC DESIGN
`NUMBER
`VOLU~E
`PAGE-BEGIN 1
`PAGE-END
`YEAR
`1974
`IIONTH
`
`16
`0
`
`CITED-BY FORREST(I974)
`CITED-POSITION 2, 3
`
`PUBLISHER
`CLASS-I/O
`
`ACADE~IC PRESS
`
`,HER ARTICLE TO BE REGISTERED BY THE SAllE AUTHOR ==>H.
`( HEXT ARTICLe =•>
`UPDATE OR DISPLAY OF CURRENT ARTICLE•=>U, DELETE•=>D. QUIT==>Q )
`
`Fig. 5 Display screen of the retrieved results.
`
`retrieval by author name is effective in the input system. The retrieval speed does not
`decrease extremely according to an increase of the registration number because inverted file
`is created with a key for the specified author field.
`The specification of 'C' to the first argument in the retrieval screen of Fig. 4 is to display
`If the content can not be inputted in the screen (over 250
`a successively retrieval screen.
`words), then to specify 'C' in the bottom argument, the continued screen for the content is
`displayed. The capacity of the content field is one thousand characters from the limit of the
`data base management system ADABAS. Fig. 7 shows the interactive retrieval screen
`except author and title.
`To input surname of author, title, journal name, publication year, key word (s), article ID,
`surname of citing author and article number (when it is plural, to specify the beginning value
`and the ending value), these retrievals are executed. When' A' is specified into ( ) of the left
`end, the intersection of those specified items is set as the retrieval condition.
`
`3. 4 Intelligent input system
`It is a weak point that we can not use the interactive input method described in 3.3 at the
`
`008
`
`Facebook Inc. Ex. 1213
`
`

`

`26
`
`Tatsuki SAITO
`
`+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++r+++++++++++++++++
`THE FOLLOWING ARTICLE ALREADY REGISTERED I I
`(MPDS5ED)
`
`++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++t++++t+++++++++++
`ARTICLE-NO
`ARTICLE-ID
`COO~S(1967l
`
`TITLE
`
`:
`
`SURFACES FOR COMPUTER AIDED DESIG~ OF SPACE FORMS
`
`COONS
`AUTHO!: 1
`APFIL. NO. 1 : AFFILIATION
`
`S. A.
`
`AFFIL. XO.
`
`JOURHL MASSACHUSETTS
`VOLUME MAC-TR-41
`PAGE-BEGIN 0
`YEAR
`1967
`
`INST. OF TECH. PROJECT MAC REP.
`NUMBER
`PAGE-END
`MONTH
`
`CITED-BY COONS(1974)
`CITED-POSITION 2
`
`PUBLISHER
`CLASS-NO
`
`( XEXT ARTICLE ==>,HER ARTICLE TO BE REGISTERED BY THE SA~E AUTHOR ==>N.
`UPDATE OR D !SPLAY OF CURREn ARTICLE==>U, DELETE==>D. QUIT==>Q )
`
`Fig. 6 Display screen of the retrieved results for the same
`author
`******************************************************************************
`*
`[MPSR5-ll]•
`* RETRIEVAL WENUE
`*
`OF REGISTERED ARTICLES
`FOR FILE
`:
`CAD
`******************************************************************************
`*
`LAST LOGON-UID
`: Xl0044
`DATE
`:
`o
`*
`LAST ARTNO = 3028
`* ========~=====~======================~==================================== *
`* ( RETRIEVAL CONDITION : OR==>
`*
`.AND==>A )
`* .......................................................................... *
`ARTICLE-NO =
`*
`*
`ARTICLE-ID =
`*
`*
`*
`*
`JOURNAL:
`*
`* •
`*
`VOLUME:
`)
`*
`*
`NUMBER:
`)
`*
`*
`KEYWORD:
`)
`*
`* • •
`KEY!!ORD:
`)
`*
`KI!YI!'ORD:
`}
`*
`KEY!!ORD:
`*
`)
`*
`KEYWORD:
`)
`*
`*
`KEY!!ORD:
`*
`)
`*
`***************************~**************************************************
`
`CLASS-NO =
`
`Fig. 7 Retrieval screen for items of key field except author
`and title.
`
`time except the service time of the on-line TSS. Then, it is necessary to input data by using
`a personal computer. An intelligent input system was developed. This system can auto(cid:173)
`matically extract the bibliographic items written according to the format as follows.
`(1) To be written is the following order and format;
`AUTHOR I AFFILIATION
`"TITLE"
`JOURNAL, VOLUME, NUMBER, BEGINNING-PAGE- ENDING-PAGE,
`(YEAR - MONTH), PUBLISHER
`#CLASSIFICATION- NUMBER
`~KEY-WORD
`(CONTENT)
`*REMARKS
`@CITED-POSITION
`
`009
`
`Facebook Inc. Ex. 1213
`
`

`

`Design and Implementation for Scientific Article Data Base
`
`27
`
`(Bibliographic items of cited articles continue here in order as described above)
`(2)
`Items except "TITLE" may be omitted.
`
`3. 5 Intelligent input system using rule system
`This input system is implemented in NATURAL by using production system like
`method. The production system consists of three parts, that is, Rule part, Data Base part
`and PSI (Production System Interpreter).
`1) Rule part is expressed as the order pair LHS-+ RHS.
`LHS (Left Hand Side) is called the left side rule or the condition part, and RHS (Right
`Hand Side) is called the right side rule, or the action part. This can be expressed as follows.
`IF { (LHS)=T } ( -+ ) THEN DO { (RHS) },
`that is, it means that if the condition part LHS is true then the action part RHS is executed.
`2) Data base part is called working memory under certain circumstances, and it is the part
`in which the string of a processing object is stored. When there is any rule that will make
`LHS true in the rules, the objective string is changed to become an another state, then it is
`called "production system".
`3) PSI (rule application part) is controlled all over the production system. PSI matches
`the rule that is stored in the data base with strings of processing object. If LHS was true,
`then RHS is applied. · As the rule generally exists in plural, if LHS is false, then the next rule
`is examined.
`Thus repetition processing is called recognize-act cycle. The production system of the
`type that is described above is called antecedent-driven production system or forward
`chaining production system. The production system of the type that requires LHS to
`execute RHS is called consequent-driven production system or backward chaining produc(cid:173)
`tion system.
`Our system is not the proper production system but production system like system
`because the data base management system ADABAS does not support AI language for the
`production system. Consequently adopting the non proper production system, could not be
`avoided but the following description is the production system because it is designed by
`production system originally.
`This system is the antecedent-driven type in case of production system. As a part of the
`rule of this system, extracting rules of author name and affiliation, extracting rules of a title,
`rules of automatic numbering of article ID and a practical processing example are mentioned
`below, and raw data is shown in Fig. 8.
`[ Extraction of author name and affiliation ]
`Rl LHSl Colon(:) exists. RHSl The string of the left side of colon is set as data of author
`information.
`R2 LHS2 Colon(:) does not exist. RHS2 An author name is set as ANON ..
`R3 LHS3 Slash(/) exists in the author information data. RHS3 Set the left side of the
`processing strings as an author name, and set the right side of the processing strings to an
`affiliation.
`
`010
`
`Facebook Inc. Ex. 1213
`
`

`

`28
`
`Tatsuki SAITO
`
`IBM RESEARCH LABORATORY:
`E.F.CODD
`I
`"A RELATIONAL MODEL OF DATA FOR LARGE SHARED DATA BANKS"
`C,ACH,13,6•377-387,(1970-6)
`IIS1-DBMS-R1
`"DATA BANK•DATA BASE.DATA STRUCTURE.DATA ORGANIZATION.HIERARCHIES OF DATA.
`NETWORKS OF DATA.RELATIONS.DERlVABILITY.REDUNDANCY,CONSISTENCY.COMPOSITION,
`JOIN.RETRIEVAL LANGUAGE,PREDICATE CALCULUS,SECURITY.DATA INTEGRITY
`(0 A .• ! RELATIONAL HODEL AND NORMAL FORH,1.1 1 .• 1.2 DATA DEPENDENCIES IN PRESENT
`SYSTEHS,1.2.1 ORDERING DEPENDENCE.1.2.2 INDEXING OEPENDENCE,1.2.3 ACCESS PATH
`DEPENDENCE.1.3 A RELATIONAL VIEW OF DATA,1.4 NORMAL FORH.1.5 SOME LINGUISTIC ASP
`ECTSo! ,6 EXPRESSZSLEoNAMED.AND STORED RELATIONS,z REDUNDANCY AND CONSISTENCY,z.1
`OPERATIONS ON RELATIONS,2.1 .1 PERMUTATION,z.1 .2 PROJECTION,2.1.3 JOIN,z.1 .4
`COHPOSITION,z.1 .5 RESTRICTION,z.z REDUNDANCY,z.2.1 STRONG REDUNDANCY,z.z.z WEAK
`REDUNDANCY.2.3 CONSISTENCY.2.4 S. ,3 AC.l
`*S
`Q1 .1
`CHILDS,Q.L.'
`"FEASIBILITY OF A SET-THEORETICAL DATA STRUCTURE-A GENERAL STRUCTURE BASED ON A
`RECONSTITUTED DEFINITION OF RELATION"
`P.IFIP C 68,162-172,<1968l,NORTH HOLLAND PUB. CO.
`Q! .1
`LEVEIN,R.E.o MARON.H.E.o
`"A COMPUTER SYSTEM FOR INFERENCE EXECUTION AND DATA RETRIEVAL"
`C,ACM,10o11,715-721. (1967-11)
`Q! .1
`BACHHAN,C,W.'
`•sOFTWARE FOR RANDOM ACCESS PROCESSING"
`OATAMATION.36-41 ,(1965-4)
`Q1 .1
`MCGEE,W.C.'
`"GENERALIZED FILE PROCESSING"
`ANNUAL REVIEW IN AUTOMATIC PROGRAMHING,So13,77-149,(1969),PERGAHON PRESS
`
`Fig. 8 An example of raw data for this intelligent input sys(cid:173)
`tem.
`
`R4 LHS4 Comma(,) exists in the author name_ RHS4 Set the left side of the processing
`strings as a surname, and set the right side of the processing strings as a name except
`surname.
`A processing example is shown in Fig. 9.
`
`OF
`SCREEN
`DISPLAY
`LAST LOGOH-UID: Al0118
`
`(MPSTl-OA]
`READ
`DATA
`EXTRACTED
`DATE: 89-02-10 TIME: 17:13:46.0 ARTHOLST: 3048
`
`ARTICLE-NO 3049
`---------------------~---------------------------------------------------------
`(SURHAJIE)
`(F l RST, Ill DDLI! HAllE)
`E. F-
`
`AUTHOR 1 CODD
`AUTHOR 2
`AUTHOR 3
`AURHOI! 4
`AUTHOR 5
`
`01
`
`AF F I L. RO.
`AF F I L. NO.
`AFF I L- NO.
`AFF IL. HO.
`AF F I L. NO.
`
`AFF I L. NO.
`
`AFFILIATION
`
`IBM RESEARCH LABORATORY
`
`AFF I L. HO. 2
`
`AFFILIATION
`
`.. "................ !HPUT RAW CONNECTED DATA
`
`(+STRIHGC) .... - ·- ·- .......... ..
`
`CODD
`
`Fig. 9 An example extracted author name and affiliation.
`
`[ Extraction of a title ]
`R5 LHS5 There are two double quotation marks ("). RHS5 Set a string surrounded by
`double quotation marks as a title_
`R6 LHS6 There is no double quotation mark. RHS6 Display a warning of NON TITLE.
`An example extracted title, etc. from the raw data of Fig. 8 is shown in Fig. 10.
`[ Numbering of article and naming of article ID ]
`
`011
`
`Facebook Inc. Ex. 1213
`
`

`

`Design and Implementation for Scientific Article Data Base
`
`29
`
`-------------------------------------------------------------------------------
`JOURllAL C.ACll
`HUliBER 6
`VOLUWE 13
`PAGE-BEGIN 377
`PAGE-END 387
`YEAR 1970
`IIOHTH 6
`---------------------------------------------------------~---------------------
`PUBL I SilER
`CLASS-KO 51-DBIIS/R
`;;~;~;~=·;~~~·;~~~··======================·;;;~=;;;~===========================
`DATA STRUCTURE
`DATA ORGANIZATION
`,
`HIERARCHIES OF DATA
`NETWORKS OF DATA
`.. "· • • " .. • .. • •..
`IKPUT RAW CONNECTED-DATA
`(+STRIIiGC) .. • .. • • • • • • • .. · · · • • • • •
`
`"DATA BANK, DATA BASE. DATA STRUCTURE, DATA ORGANIZATION, HIERARCHIES OF D
`ATA, NETWORKS OF DATA,RELATJONS.DERIVABILITY,REDUNDANCY,COHSISTEHCY (0
`A., 1 RELATIONAL IIODEL AND NORIIAL FORI!, 1.1 I. ,1. 2 DATA DEPENDENCIES
`IN
`PRESENT
`
`*
`
`Fig. 10 An example extracted title, etc.
`
`R7 LHS7 An article number (ARTNO) has not been registered. RHS7 Add one to the
`final article number (ARTNOLST).
`R8 LHS8 The time when the article was published is stated. RHS8 Set article ID
`(ARTID) = SURNAME (YEAR).
`R9 LHS9 The time when the article was published is not stated. RHS9 Set the YEAR =
`0000, that is, article ID (ARTID) = SURNAME (0000).
`RIO LHSlO Article ID (ARTID) agrees with a registered article ID. RHSlO Display a
`warning of article ID duplication.
`Rll LHSll Article ID (ARTID) conforms with a registered article ID and the published
`year is the same and the title does not conform. RHSll N arne the new article ID that is
`changed the YEAR part by the alphabetically consequential character as follows:
`0 the last article ID = XXXXXXXX (YYYYB)
`!
`0 the next article ID by the same author and published at the same year
`XXXXXXXX (YYYYC)
`
`4. Generating System of Relation Matrix
`
`This system is to set up a matrix of the relation among scientific articles by user's
`specification. A rowwise number and a columnwise number of the matrix mean an article
`If it is for the citation relation matrix, then the matrix that ( i , j )element is 1 when
`number.
`article i is cited by article j is outputted.
`It is the asymmetric matrix because the citation
`relation is not reciprocal with regard to time.
`If it is for keyword relation matrix (or title relation matrix), then the matrix that (i , j)
`element is m when there are m common keywords (or terms) between article i and article
`j is outputted.
`It is the symmetric matrix. If it is for author relation matrix, then the matrix that (i ,
`j)element is 1 when an author wrote article i and article j commonly is outputted. Also, it
`is the symmetric matrix.
`
`012
`
`Facebook Inc. Ex. 1213
`
`

`

`30
`
`Tatsuki SArTo
`
`5. Exploratory Illustration
`
`The source data list inputted in data base system ANGEL is shown in Table 1 summarily.
`In the field for computational three dimensional geometry mainly (the research field to be
`picked up is expressed in "CONTENTS" item), including both articles to cite and articles to
`be cited, about 5,000 scientific articles of CAD (Computer Aided Design)/CAM (Computer
`Aided Manufacturing) and of a few data base fields are acquired into the data base.
`Hereinafter, in regard to the relationship between a citing article and a cited article, several
`considerations are taken up in more detail.
`
`Bibliog. (FILE NAME)
`
`CONTENTS
`
`To cite I To be cited
`
`Cards.
`
`Computer Aided Geometric Design (CAGD)
`PRO LAMA T' 69 (P69)
`PRO LAMA T' 73 (P73)
`PRO LAMA T' 76 (P76)
`JICST {]!CST)
`Computer Aided Design (CAD)
`IEEE Tranactions on Computers (IEEETC)
`Proceedings of IEEE (PIEEE)
`Design Engineering Projects. (DEP)
`
`SIAM Journal on Control and Optimization (SIAM)
`Computer Journal (CJ)
`Communication of ACM (CACM)
`Journal of ACM (JACM)
`Journal of Mathematics and Physics (JAPSAM)
`Journal of Approximation Theory (JAT)
`Numerical Control in Manufacturing (NCM)
`Proceedings of Royal Society of London (PRSL)
`Journal of Mathematics and Mechanics (JMM)
`Journal of Approximation Theory (JAT)
`Journal of Mathematical Physics (JMP)
`Numerische Mathematik (NM)
`
`total
`
`CAGD
`CAD I CAM
`CAD I CAM
`CAD I CAM
`CAD
`CAD
`general
`imag. process. patt. recog.
`CAD I CAM
`math. prog.
`general
`comp. graph.
`c