應用記憶體內運算於多維度多顆粒度資料探勘之研究―以醫療服務創新為例 - 政大學術集成

全文

(1)。. （. （. ―. A Research Into In-memory Computing In Multidimensional, Multi-granularity Data Mining ― With Healthcare Services Innovation. ○.

(2) . .

(3) （. （. (knowledge) (. ). → （. （. Hadoop MapReduce (single-node). （ MapReduce Spark. ○ （、. Map. 「. （. Reduce. （. （ 100. 1. Spark. Hadoop. 4. Spark. Hadoop Spark. （. （. Apriori Algorithm.

(4) Abstract Under the population aging and population growth and rising demand for Healthcare. Healthcare is facing a big issue how to effectively deal with huge amounts of data. Cased by new healthcare services or applications (such as electronic health records or health care, etc), and also medical institutions in accordance with government policy for long-term preservation of a large number of patient data. But the traditional algorithms for mining association rules, subject to considerable restrictions on their effectiveness. Therefore, many studies suggest that the association rules algorithm in a distributed computing, such as Hadoop MapReduce framework implements parallel to process huge amounts of data operations. But in fact, MapReduce does not apply to require intensive iterative computation algorithm of association rules. Studied in this Spark in-memory computing framework, implemented on a distributed cluster parallel mining association rules mining multidimensional granularity, the experimental results can be summed up in the following three points. 1th, when data is small, due to the parallel data flow consists of Map and Reduce, so not much in the small-scale processing of benefits. 2nd, when the data size is large, parallel strategy models and stand-alone obviously significant differences overall running time is 100 times as much when the item number is greater than 10,000, however, stand-alone version cannot run due to insufficient memory, but parallel strategies can still run. 3rd, overall Spark though somewhat slower than the single version in small scale processing speed, but the running time is less than 4 times times the Hadoop. Massive processing speed Spark is still superior to the Hadoop version. Therefore, when working with large data, operational efficiency and expansion elasticity, Spark for optimum solutions.. Keywords: Data Mining, Multidimensional and Multi-granularity Data Mining Association Analysis, In-Memory Computing, Apriori Algorithm.

(5) .................................................................................................................................................. I .................................................................................................................................................. I ............................................................................................................................................ IV ............................................................................................................................................ VI. ........................................................................................................................... 1. 1. ............................................................................................................................... 1. 2. ............................................................................................................................... 2. 3. ............................................................................................................................... 3. 1. 2. ................................................................................................................... 4 ........................................................................................... 4. （. ............................................................................................................................... 5. 2.1. .................................................................................................................. 5. 2.2. .............................................................................................................. 6. ................................................................................................................. 11. 3.1. QUANTITATIVE ASSOCIATION RULES ............................................................................................. 11. 3.2. ASSOCIATION RULE CLUSTERING SYSTEM ...................................................................................... 12. 3.3. BOOLEAN MATRIX BASED APPROACH ........................................................................................... 12. MAP-REDUCE .......................................................................................................................... 13. 3. 4. 4.1. MAP-REDUCE. ..................................................................................................... 13. 4.2. APACHE HADOOP ..................................................................................................................... 16. ( 1 ).. HADOOP MAPREDUCE .............................................................................................................. 16. ( 2 ).. APACHE HADOOP HDFS ............................................................................................................ 17 i.

(6) . (IN-MEMORY COMPUTING) ............................................................................... 18. 4.1. ................................................................................................. 18. 4.2. APACHE SPARK ......................................................................................................................... 19. 5. 1. ................................................................................... 22. ............................................................................................................................. 22. 1.1. （. (MULTI-DIMENSION DATABASE, MD) ................................................................... 22. 1.2. (CONCEPT HIERARCHY, CH) ............................................................................ 23. 1.3. (MULTI-DIMENSION PATTERNS) .............................................................................. 24. 1.4. (MULTI-DIMENSION ASSOCIATION RULES) .......................................................... 26. 2. 「. ......................................................................................................................... 27. 2.1. 0. ...................................................................................... 29. 2.2. 1. .......................................................................... 30. 2.3. 2. .......................................................................... 31. 2.4. 3. .......................................................................... 33. 3. ................................................................. 34. 3.1. ........................................................................................................ 35. APRIORI. 3.2. ............................................. ERROR! BOOKMARK NOT DEFINED.. ...................................................................................................... 37. 1. ................................................................................................. 37. 2. ................................................................................................. 38. 2.1. ............................................................................................................. 39. 2.2. ......................................................................................................... 40. 3. 3.1. →. （. .................................................................... 42. ............................................................................................................................. 42 ii.

(7) 3.2. 「. 3.3. ............................................................................................................................. 43 ............................................................................................................................. 44. ...................................................................................................... 46. 1. ......................................................................................................................... 46. 2. ..................................................................................................................... 46 ........................................................................................................................................ 47. iii.

(8) . ..................................................................................................... 7. AIS. ................................................................................................ 7. APRIORI. APRIORI. (. (. MAPREDUCE. (. （. ). （. .............................................................................................. 12. ............................................................................................................................. 12 ) ARCS. ( 「. HDFS. 2 ...................................................................................................................... 11. （ )：. : AGRAWAL,1994) .................................................................................... 8. (（. (. ....................................................................... 13. ). (JEFFREY DEAN AND SANJAY GHEMAWAT,2004) .................................................. 14 : HTTP://CCWEB.KM.NCCU.EDU.TW) .................................................................. 17 ......................................................... 19. （. (ZAHARIA ET AL.,2012) ....................................................................................... 20. SPARK. RDD. (ZAHARIA ET AL.,2012) ................................................................................. 21. SPARK. (ZAHARIA ET AL.,2012) ................................................................................ 21. .......................................................................................... 23. CH1. ........................................................................................... 24. ........................................................................................... 24. ..................................................................................................................... 25. ......................................................................................................................... 25. ......................................................................................................................... 26. ................................................................................................. 27. ................................................................................................................. 28. ................................................................................................................................. 28. （. ..................................................................... 29. MD ： iv.

(9) . ................................................................................................. 29. (MAPPING TABLE) .............................................................. 30. ............................................................................................................. 31. . ............................................................................................................. 32. ..................................................................................................... 33. ................................................................................................... 32. ............................................................................................................. 34. 「「. QIU, ET AL, 2014) .............................................................................. 35. 1 (（. QIU, ET AL, 2014) ........................................................................... 36. 2 (（. ..................................................................................................... 36. ......................................................................................................... 38. TYPEI. -. .................................................................. 39. TYPEII. -. ................................................................. 40. TYPEIII. -. ................................................................. 41. TYPEIIV. ................................................................... 41. -. ................................................................................................................. 42. v.

(10) 1. 1 ...................................................................................................................... 5. （. 2. （. MD .................................................................................................................... 22 ............................................................................................................................... 37. 3 （ 4. ........................................................................................................................... 37. 5. ................................................................................................................................... 38. 6. （. MD ........................................................................................................................ 42. 7. (ELEMENT PATTERNS, EI) .......................................................................................... 43. 8. (MAPPING TABLE) ....................................................................................... 44. 9. ................................................................................................... 45. 10. 0.5. 1. ...................................................................................... 46. vi.

(11) 1 ； IDC. （. 2010. 2020. 81% （ IDC. （. （. 2014 48%. (. ). →. （（（. （. (knowledge) （） (QoS) （. →. （。（ → →. (Korea Medical Insurance Corporation ,KMIC). (Chae et al., 2001) 48. 43. （、. （（. 1.

(12) 2 (Agrawl et al., 1993). （. (Large itemsets) (Association rules). (iterative). 」. k-1. （. I/O. （. (. ). (. ). （ 2009. Google. (Cloud Computing). （. (Task). (Process) MapReduce. － (Dean and Ghemawat,2004). （ Hadoop MapReduce. (Yang X.Y., 2010; Li L. & Zhang M., 2011; Li N et al., 2012) (single-node). MapReduce (Dean and Ghemawat,2004). MapReduce. MapReduce. I/O. 2. (job). HDFS.

(13) 3 Apache Spark (Matei Zaharia et al.,. ； 2012). Spark. （ In-Memory Computing (Random Access Memory. （ RAM). （ Hadoop MapReduce Apache Spark Spark Hadoop MapReduce. 18. Apriori. (Hongjian Qiu et al., 2014) (Chiang,. （ Johannes, and Chia-Chi Wu,2005) (patterns). （. (. ). (. ). Spark. （. →. 3. （.

(14) 1. （. （. ）. (QoS). →. （. 。. （. （. :. → (1). : 。. → 。. (2). :. （ VIP. (3). : 。. (4). : MRI. (5). →. / 。. :. （. 4.

(15) 2 (Association Rule). （. AIS. （ (scalability). （ Apriori. (Agrawl et al., 1994). AIS. Apriori. （. （ 1. {. à. （. }. (Cross-Selling) TID. ITEMS 1. {. ,. }. 2. {. ,. ,. ,. }. 3. {. ,. ,. ,. }. 4. {. ,. ,. ,. }. 5. {. ,. ,. ,. }. 1. 1. （. 2.1 （. D (. 1. ). TID. （. I = #$ , #& , ⋯ , #( *+. I. T = *$ , *& , ⋯ , *-. (itemset). X=. I. A⟶B ( 2. 3 = ∅ ). A⟶B. 5. B. A. B A.

(16) A. （. B. (A) = . (A ⟶ B ) = . A. A. A. D. （（. D. B. D. （. A. D. （. D. 2.2 (1). (large. itemsets or frequent itemsets). (minsup) (2). (1). (minconf). AIS (1). （. 2n. n. Apriori. AIS Apriori. AIS. Apriori. 「. k-1. k (. ). 6.

(17) AIS. Apriori. 7.

(18) Apriori. (（. : Agrawal,1994). Apriori. 」（. （. Apriori. （. l. AprioriTid （ TID （ TID （. (Agrawal et. al.,1994) l. FP-TREE prefix-tree. （. divide-and-conquer （ (Han et al.,2000). 8. 2.

(19) （. l. Partition. n. （. n. l. Sampling. （. （. （（. (Chuang et al.,2005) l. Segmentation. （ (patterns). (Chiang, Johannes, and Chia-Chi Wu, 2005). l. FDM （ Apriori. （. distributed (global) pruning (Cheung et al., 1996). DDM. local pruning. (Schuster. et al.,2004) l. Apriori 9. MapReduce.

(20) Division. （ Allocation. （. (Yang X.Y., 2010; Li L.. & Zhang M., 2011; Li N et al., 2012) l. YAFIM Spark. Apriori RDD. (In-memory. computing). (Qiu, et. al, 2014). l. dataset filtering （（ (Wojciechowski et al, 2002). l. Category-based. Apriori. (Tien Dung Do et al, 2003) l. RARM. （. （. （（. (Das et. 10.

(21) al, 2001). 3 (single-dimensional association rules) (". "). (". "). (multidimensional association rules) ". "20" ) ⋀ (. (. ". ")⇒(. ,. ") (intra-dimensional. association rules) (hybrid-dimensional association rule). 3.1. （. ： (. （. Quantitative Association Rules. )（. （. (. （（（. ：. (Ramakrishnan Srikant, Rakesh Agrawal, 1996). （. 11. 2. ).

(22) (. （. (. ). （. 3.2. Association Rule Clustering System. ARCS. LHS (. （（. )：. ). (. 2 ). (. ). (BinArray). ：. (. 3.3. RHS. （. ). (Brian Lenty, et al., 1993). Boolean Matrix based Approach. (Chiang, Johannes, and Chia-Chi Wu, 2005) (Neelu Khare et al., 2010) （ n. (. *（. m. ”1”. （ (. 2:×<. ：. （ ”0”. ) ). A′ m. 12. k+1 2:>×<.

(23) (. 4. ) ARCS. (. ). Map-Reduce. Google. (Google File System, GFS)(Sanjay. Ghemawat et al.,2003). MapReduce. （. Sanjay Ghemawat,2004). GFS. BigTable. (Cloud Computing). ，. 4.1. (Jeffrey Dean and. （（. （ (Big Data). Map-Reduce. MapReduce. MapReduce. （（. MapReduce. MapReduce （. （. MapReduce. （ Map. (. （. （. Reduce. （. ). MapReduce. 「. 「 (1) MapReduce. （ 13. M. 「.

(24) 16~64M. (. ). MapReduce. Master (2) Master. worker M. Map. worker. worker. (3). Map （. worker. （. key/value pair. Map. key/value pair. Map. MapReduce. (4). 「. Map （. Reduce. worker. （ Map. R. (Jeffrey Dean and Sanjay Ghemawat,2004). worker. （. key/value pair. （ 14.

(25) Map. Map. key/value pair. Map (5) partition. R. （. Master (6). Master. Master Reduce. Reduce. （. worker. worker. （. Map worker. （. ，. key. （. (7) Reduce worker. key. Reduce. value. Reduce Reduce. (8). Map. Reduce. Master. MapReduce R. RPC. Reduce worker. key. （. Reduce. R. R （. MapReduce. 15. worker. （.

(26) 4.2. Apache Hadoop. Apache Hadoop Google. Apache Lucene. MapReduce. Doug Cutting. Java Apache Hadoop. Hadoop Common. Hadoop Distributed File System (HDFS™) Hadoop YARN Hadoop MapReduce Hadoop MapReduce. HDFS. (Apache. Hadoop,2015) ( 1 ).. Hadoop MapReduce. Hadoop MapReduce JobTracker. JAVA. Map Reduce. JobTracker. TaskTracker. （ TaskTracker. (. ) key-value. （. Hadoop MapReduce l. （. l. （. TaskTracker (. TaskTracker. ). l. Hadoop MapReduce. 16.

(27) ( 2 ).. Apache Hadoop HDFS. HDFS (Hadoop Distributed File System) Master. NameNode. Name Node. (master/slave) slave (. （ (open/close/create... DataNode. ). Data Node Data Node. （ HDFS l. HDFS. （. （. : （. （. l. l. （. DataNode. （. （. （. （. ，. （. （. （. （. （ DataNode. HDFS. (. : http://ccweb.km.nccu.edu.tw) 17. （. ).

(28) 5. (In-Memory Computing). 4.1 （. （ I/O. CPU. 「. （. 、. （. ；. （. （. (Random Access Memory RAM). （. I/O. （（（「（. （. 「. (Data Warehouse)（. （. （. － (Data Mart)（. （. (Data Mining). （ OLAP (Online Analytical Processing) Web Portal. （. （ (in-memory enabled). （. （. 「（. （. I/O. （. （ Spark. （ Pregel. GraphLab. Mammoth. Phoenix. GridGain. Storm Yahoo! S4 Spark Streaming 18. (. Piccolo 「. MapReduce Online. ).

(29) （ (（. 4.2. : In-memory Computing Technology The Holy Grail Of - Deloitte). Apache Spark. 2012. (Zaharia et al.,2012). Spark （. (Resilient Distributed Datasets ,RDDs) l. :. （. （. (iterative algorithms). (interactive data mining). （. （. l. lineage. RDD ：. (transformations). (actions). （. lineage （ RDDs. MapReduce (reuse). （. (sharing) （ 19. (. ,.

(30) distributed file system). I/O. （ spark. spark. 1TB （. 5-7. Spark. RDD. Java. Python, Scala,. API HFDS. RDD RDD. （ (worker). lineage HDFS. 「. spark. Spark ：. (Zaharia et al.,2012). RDD. (transformations). RDD. （. RDD. RDD. （ RDD). l. 「 RDD. （. l. 20. (actions). map. filter. flatMap. RDD ：. count collect. (. RDD ：（ RDD. (Dependency). ： Narrow Dependency. 20. (. ：. (lazy). reduce). (action). ：. （.

(31) Wide Dependency. RDD. RDD. RDD. RDD. RDD. (Zaharia et al.,2012). RDD. ：. RDD ： RDD. RDD RDD. RDD RDD. 「 RDD. ：. (lineage) RDD narrow. 「. lineage graph. ，. RDD. RDD. (. spark. (Zaharia et al.,2012) 21. DAG. ).

(32) (Chiang, Johannes, and. （ Chia-Chi Wu,2005) 「. Spark. （. 1 （ (concept hierarchy, CH). (multi-dimension database, MD) MD. MD. (multi-dimension patterns) (multi-dimension rules). 1.1. (Multi-dimension Database, MD). （（. 2. （. （（. （. MD 2. (. ). ○. (. ). T_ID 1. 2012/3. 20. 2. 2012/3. 44. 3. 2012/4. 18. 4. 2012/5. 33. 5. 2012/6. 60 2. （. 22. MD. (. ).

(33) 1.2. (Concept Hierarchy, CH). dimx. MD. （. x ”. CHx. ”. dimx. dim1. CH1 CHx. (root) MD. CHx. CH1. CHx l l. (root). CHx. (leaves nodes) da(x,s). CHx. (dimension atom, da). CHx. s. 2005 ”. ”. 12. {2005 Jun, 2005,Feb,…,2005 Dec} = { dc(x,1), dc(x,2) ,…, dc(x,12) }. l. (non-leaf nodes) dc(x,t). (dimension compound, dc) CHx 23. t.

(34) dc(x,1) = {2005 Spring} 2005May}. CHx. 1.3 MD. {2005 March, 2005Apr,. (Multi-Dimension Patterns) n. {CH1, CH2, CH3, …, CHn}. n. CH. (multi-dimension pattern, p) <p1, p2, p3,…,pn>. pk. dimk. CHk (element patterns, Ei). (generalized patterns, Gj). p. 1.2. p. da dc. da={branch1, brach2, web}. dc={Taipei, Any} da={Male, Female}. 24.

(35) dc={Any} Female>. p Ei. Gj. da. <branch1, Any>. dc={Any}. p Ei. <web,. MD. （. MD. p=. (element segmentations, T[Ei]) p=Gj. (Combination. segmentations, T[Gj]). T[Gj] T[Gj] (Any, Female). Female),(B02, Female),(web, Female)}. 25. {T[Ei] = (B01,.

(36) 1.4. (Multi-dimension association rules) (P,r). 1.3. P. r (6). (Full match) (minsup). MD. （ (minconf). T[Gj]. T[Gj]. T[Ei]. T[Gj] = {2004 Spring}. 2004Mar, 2004 May}. 3. T[Ei]={2004 Apr,. T[Gj]. (A→B). T[Ei]. (A→B) (7). (Relaxed match) (minsup). （ (minconf). MD (match ratio, m). 0<A≤1. T[Gj] T[Ei] (. m /. ). 26.

(37) 2. 「「 MD. （ (minsup),. (minconf),. (match ratio, m) n. 「 MD. (. （. CHn. ： (mapping table). ，. ). T[Ei]. R Ei. T[Gj]. R Gj. (P,r) T[Ei] R Ei. T[Gj]. 1.4. R Gj T[Gj]. T[Ei]. T[Gj] <2004 spring, Branch01, Student, Any> and <2004 spring,. Branch01,Any,Male>. <2004 spring, Branch01, Student, Male>. 27.

(38) 1) Input: 2) Multi-dimension transaction database: MD; 3) concept hierarchies for each dimension: CHx (x = 1 to n); 4) user define threshold: minsup, minconf, match ratio m; 5) Procedure: 6) Phase0: 7) generate all Ei and Gj by CHx (x = 1 to n); 8) build the pattern table; 9) Phase1: 10) for all Ei 11) discover all association rules r in T[Ei] as REi; 12) Phase2: 13) for all Ei 14) for all Gj that Ei ⊂ Gj 15) update RGj using REi; 16) Phase3: 17) for all Gj 18) for all r (which satisfy m) in RGj 19) output (Gj, r); 20) Output: 21) All multi-dimension association rules (p, r);. 28.

(39) 2.1. 0 （. )： p. CHn. CHn. n. (. MD (mapping table). ， MD. ：. Any. Spring (Any). Mar. MD. Apr. May. （. CH. Male. Female. MD ：. CH. 12. Ei = {<Mar, Male>, <Mar, Female>, <Apr, Male>, <Apr, Female >,<May, Male>,< May , Female > }. Gj = {<Mar, Any>, <. Apr, Any >, < May, Any >, <Spring, Male >,< Spring, Female >,<Spring(Any)>}. 29.

(40) 12. ，： Ei. Gj Ei. ”0”. Gj. Gj (Mar) (Apr). ”1”. (. (May) (Spring, Male). ). Ei. (Spring, Female) (Spring). (Mar, Male). 1. 0. 0. 1. 0. 1. (Mar, Female). 1. 0. 0. 0. 1. 1. (Apr, Male). 0. 1. 0. 1. 0. 1. (Apr, Female). 0. 1. 0. 0. 1. 1. (May, Male). 0. 0. 1. 1. 0. 1. (May, Female). 0. 0. 1. 0. 1. 1. (mapping table). 2.2. 1 p. T[Ei]. T[Gj]. T[Ei]. R. T[Gj]. Ei. R. （. Gj. Ei = {<Mar, Male>, <Mar, Female>, <Apr, Male>, <Apr, Female >,<May, Male>,< May , Female > } R. T[Ei] R. Ei. Gj. R Ei Apriori Apriori. (Agrawl et al., 1993) Spark. （. Aprioir. 3. 30.

(41) 2.3. 2. ”bottom to top” R. R Ei.. R Gj. 1. T[Ei]. 0. Ei. T[Ei]. T[Gj]. (1). R Gj. (Full match). R Gj R Ei R Gj. T[Gj]= (Spring, Male) T[Ei] = {(March, Male) (Apr, Male) (May, Male)} R Ei 7 Male). T[Ei] = (March, Male). R Ei. T[Ei] = (Apr,. R Ei. {C→A} {B→C} {AB→C}. T[Ei] = (Apr, Male). 4 T[Ei] = (May, Male). 3 (Spring, Male). {A→B} {A→C} {BC→A} R Gj. 31. T[Gj]=.

(42) (2). (Full match). R Gj R Ei. －. (match ratio, m) R Gj 1) for all REi 2) for all Gj ⊃ Ei 3) for all r in REi 4) if (r ∉ RGj) 5) add r to RGj; 6) RGj.r.count = 1; 7) else 8) RGj.r.count++;. T[Gj]= (Spring, Male) T[Ei] = {(March, Male) (Apr, Male) (May, Male)} T[Ei] = {(March, Male)} ). ( m 32. =. T[Ei] = {(March, Male)}. /.

(43) m=0.6667(. {AC→B}. 1/1 = 1). T[Gj]= (Apr, Male). m. T[Gj]= (May, Male) 1. 1/3=0.333 1. 0.667. 5. {AC→B} {BC→A}. 2.4. {A→B} {A→C} {AB→C}. T[Gj]= (Spring, Male). 3 (Ei,r). Gj. R Gj. r. (Gj,r). T[Ei]. Gj. T[Ei]. R. Spring. Ei.. Apr 33. R. Gj. May. r.

(44) 3 Apriori. （. Apriori. Spark. （ Apriori. (Qiu, et al, 2014). Apriori 」. （ 3. 3. （ <. （ <key,. （. key. 1>. value key. >. spark RDD. （. （ 1. 「 (k+1). 34. 「.

(45) 3.1 (1) 「. Apriori. 1. 「. 1(. （. ：. ). flatMap(). spark RDD. map(). key. 2 「. k+1. < item,1> < item,count >. 1 (（. Qiu, et al, 2014). (k+1) 2(. ) Ck+1. k. Lk<item,count>. Ck+1. hash tree. RDD （：. value. key. 「. (2) 「. （. flatMap(). ：. reduceByKey(). HDFS. key, value. flatMap(). hash tree < item,1>. key. 35. map() reduceByKey().

(46) 「. 2 (（. Qiu, et al, 2014). 3.2 3.1. Apriori HaspMap. CountByValue. 36. ID. <GID,Itemsets>.

(47) (Chiang, Johannes, and Chia-Chi Wu,2005). Hadoop. Spark. 1 OpenStack. 20. 32GB. 1. 5. （. 3. Ubuntu12.04 Apache Spark 14.01 *1. CPU: 4 cores / Memory: 8GB. *5. CPU: 4 cores / Memory: 8GB. Apache Hadoop 2.7.0. 3 （. （. (datasets). （ (. (Frequent Itemset Mining Dataset Repository) （. 4) （. I. MushRoom.txt. 119. 8,124. II. Accidents.txt. 468. 340,183. III. Pumsb_star.txt. 2,113. 49,046. IBM Almaden Quest. （. （ UCI Machine Learning Repository （. IIV. Retail.txt. 16,469. 88,162. （. 4 37.

(48) 2. 4. 6. 1. TypeI -hadoop -spark. 0.25 6.08 39.83 9.66. 0.35 2.07 39.69 8.19. 0.65 0.77 39.24 7.40. 0.85 0.61 38.97 7.27 (minsup/minsec). -hadoop -spark. 0.25 4800.06 72.37 65.79. 0.35 3600.06 70.92 64.47. 0.65 163.77 64.90 31.5. 0.85 53.89 62.93 25.17 (minsup/minsec). -hadoop -spark. 0.25 20.60 217.89 54.47. 0.35 17.40 157.66 39.41. 0.65 12.42 45.65 12.042. 0.85 6.35 44.91 10.94 (minsup/minsec). TypeII. TypeIII. TypeIIV -hadoop -spark. 0.25 0.35 0.65 0.85 OutOfMemoey OutOfMemoey OutOfMemoey OutOfMemoey 42.58 42.65 43.63 42.34 24.69 24.61 22.70 22.37 (minsup/minsec) 5 38.

(49) 2.1 I. II. 1. （. TypeI （. (Map （. Reduce （. ). 10. （. Spark. TypeII. Hadoop. TypeI. Hadoop. Spark 0.65. Spark. 50 45 39.83. 40. 39.69. 39.24. 38.97. 35 30 25 20 15 10. 9.66. 8.19. 6.08. 5. 7.40. 2.07. 0.77. 7.27 0.61. 0 0.25. 0.35. 0.65 -hadoop. TYPEI. 39. 0.85 -spark. Hadoop 0.85.

(50) 100.00. 80.00. 60.00 -hadoop 40.00. -spark. 20.00. 0.00 0.25. 0.35. 0.65. 0.85. TYPEII. -. 2.2 I. II 1 1. 、. （ 1 Spark. Spark. Hadoop. Hadoop. （. TypeIII. Hadoop. Spark 0.35. Hadoop 0.25. －. 100. Hadoop. 4. Saprk 200. Hadoop Spark. Hadoop 40. I/O. 39 Spark.

(51) 100.00 90.00 80.00 70.00 60.00 50.00. -hadoop. 40.00. -spark. 30.00 20.00 10.00 0.00 0.25. 0.35. 0.65. TYPEIII. 0.85 -. -spark. -hadoop. 43.63. 42.65. 42.58. 24.69. 24.61. 0 0.25. 42.34. 22.70. 0. 22.37. 0. 0.35. 0.65. TYPEIIV. 41. 0 0.85.

(52) 3. →. （. 3.1 →. 2005. （. R201_2009~R207_2009 582,070. ○ (R201_CD2009). 。 167.09(MB). （. CD. 6. (CURE_ITEM_NO1, CURE_ITEM_NO2,CURE_ITEM_NO3, CURE_ITEM_NO4, FUNC_DATE, ID_BIRTHDAY, ID_SEX). （ (MD). （. 6. CURE_ITEM T_ID. FUNC_DATE. ID_BIRTHDAY. 1. 2005/3. 58. 2. 2005/3. 23. 3. 2005/4. 45. 15. 47. 52. 4. 2005/5. 64. A1. A5. A6. 5. 2005/6. 40 6. MD. ID_SEX. CURE_ITEM D8 10. 4. 9. 23. G9 （. 42. 44. 9. MD. CH1=. ：. 12. CH2=. CH3=.

(53) 3.2. 「 1. =1(. = 0.1. ) CH1=. 「. CH2=. CH3=. ，. (mapping table) 32. (element patterns, Ei). 43. (generalized patterns, Gj). Ei. Ei. E1. 15-24. E17. 45-64. E2. 15-24. E18. 45-64. E3. 15-24. E19. 45-64. E4. 15-24. E20. 45-64. E5. 15-24. E21. 45-64. E6. 15-24. E22. 45-64. E7. 15-24. E23. 45-64. E8. 15-24. E24. 45-64. E9. 25-44. E25. 65. E10. 25-44. E26. 65. E11. 25-44. E27. 65. E12. 25-44. E28. 65. E13. 25-44. E29. 65. E14. 25-44. E30. 65. E15. 25-44. E31. 65. E16. 25-44. E32. 65. 7. (element patterns, Ei). E1. 1. 2. 6. 8. 13. 21. 41. E17. 1. 2. 6. 10. 13. 29. E2. 1. 3. 6. 8. 14. 22. 41. E18. 1. 3. 6. 10. 14. 30. E3. 1. 4. 6. 8. 15. 23. 41. E19. 1. 4. 6. 10. 15. 31. E4. 1. 5. 6. 8. 16. 24. 41. E20. 1. 5. 6. 10. 16. 32. E5. 1. 2. 7. 8. 17. 21. 42. E21. 1. 2. 7. 10. 17. 29. E6. 1. 3. 7. 8. 18. 22. 42. E22. 1. 3. 7. 10. 18. 30. E7. 1. 4. 7. 8. 19. 23. 42. E23. 1. 4. 7. 10. 19. 31. E8. 1. 5. 7. 8. 20. 24. 42. E24. 1. 5. 7. 10. 20. 32. 43.

(54) E9. 1. 2. 6. 9. 13. 25. 43. E25. 1. 2. 6. 11. 13. E10. 1. 3. 6. 9. 14. 26. 43. E26. 1. 3. 6. 11. 14. E11. 1. 4. 6. 9. 15. 27. 43. E27. 1. 4. 6. 11. 15. E12. 1. 5. 6. 9. 16. 28. 43. E28. 1. 5. 6. 11. 16. E13. 1. 2. 7. 9. 17. 25. 43. E29. 1. 2. 7. 11. 17. E14. 1. 3. 7. 9. 18. 26. 43. E30. 1. 3. 7. 11. 18. E15. 1. 4. 7. 9. 19. 27. 43. E31. 1. 4. 7. 11. 19. E16. 1. 5. 7. 9. 20. 28. 43. E32. 1. 5. 7. 11. 20. 8. (mapping table). ( 32. 7). 32. 3.3 Spark. 11.448 9. (element patterns, Ei). (. ). 。. E0. [[P8], [C4], [47], [82], [P3]]. E1. []. E10. [[2], [19], [P8]]. E11. [[47], [P3], [P8]]. E12. [[P4], [P4, P8], [47], [P3], [P3, P8], [C4], [P8]]. E13. [[P4], [P4, P8], [47], [P3], [P3, P8], [C4], [P8]]. E14. [[P3], [P3, P8], [P4], [P4, P8], [C4], [P8]]. E15. [[C3], [P3], [C4], [P4], [P4, P8], [P8]]. E16. [[1], [1, 19], [1, 2], [7], [11], [19], [19, 2], [47], [2]]. E17. [[1], [1, 19], [1, 2], [7], [11], [19], [19, 2], [47], [2]]. E18. [[1], [1, 19], [1, 2], [7], [47], [19], [19, 2], [11], [2]]. E19. [[1], [9], [19], [47], [2]]. E20. [[1], [9], [19], [47], [2]]. E21. [[1], [9], [19], [47], [2]]. E22. [[1], [9], [19], [47], [2]] 44.

(55) E23. [[1], [9], [19], [47], [2]]. E24. [[11], [14], [9], [1], [1, 2], [19], [2]]. E25. [[11], [14], [9], [1], [1, 2], [19], [2]]. E26. [[11], [14], [9], [1], [1, 2], [19], [2]]. E27. [[11], [14], [9], [1], [1, 2], [19], [2]]. E28. [[11], [11, 2], [47], [9], [1], [1, 2], [19], [19, 2], [14], [2]]. E29. [[11], [11, 2], [47], [9], [1], [1, 2], [19], [19, 2], [2]]. E3. [[13], [13, 47], [60], [82], [P8], [P8, 40], [4], [P3], [P3, 40], [47], [C4], [C4, 60], [40]]. E30. [[19], [19, 2], [47], [9], [1], [1, 2], [11], [2]]. E31. [[19], [19, 2], [47], [9], [1], [1, 2], [11], [2]]. E4. [[82], [A6], [47], [P8], [C4], [P3]]. E5. [[82], [A6], [P8], [P8, P3], [P4], [P4, P8], [P3]]. E6. [[P4], [P4, P8], [P3], [P3, P8], [C4], [P8]]. E7. [[A6], [P8], [P8, P3], [82], [P3]]. E8. [[P4], [2], [A6], [P8]]. E9. [[A6], [47], [P4], [P8]] 9. 10. G1~G10 1(. 0.5. ). G8~G10. G1~G10. (m=1) G8:[[P8]] G9:[[47], [2], [1], [19]] G10:[[9], [1, 2], [11], [2], [1], [19]] (m=0.5) G1:[[47], [2], [1], [P8], [19]] G2:[[47], [2], [1], [19]] G3:[[2], [1], [19]] G4:[[9], [P3], [47], [2], [1], [P8], [19]] G5:[[47], [2], [1], [19]] G6:[[9], [P3], [47], [2], [1], [P8], [19]] G7:[[P3], [P8], [C4], [82]] 45.

(56) G8:[[P3], [47], [P8], [C4], [P4, P8], [P4]] G9:[[9], [47], [2], [1], [19]] G10:[[9], [1, 2], [14], [11], [47], [19, 2], [2], [1], [19]] 10. 0.5. 3 ]. [1. 1. ( [1. , 2. , 19. ]). ]. [2. , 19. 45~65. 70%. 1 Spark ○ （、. Map. 「. （. Reduce. （. （ 100. 1. Spark 4. Hadoop Spark. Hadoop Spark. （ Spark. 2 （. 46. Spark.

(57) [1] Agrawal, Rakesh, and Ramakrishnan Srikant. "Fast algorithms for mining association rules." Proc. 20th int. conf. very large data bases, VLDB. Vol. 1215. 1994. [2] Agrawal, Rakesh, Tomasz Imieliński, and Arun Swami. "Mining association rules between sets of items in large databases." ACM SIGMOD Record. Vol. 22. No. 2. ACM, 1993. [3] Apache Hadoop, https://hadoop.apache.org/,2015 [4] Chae, Young Moon, et al. "Data mining approach to policy analysis in a health insurance domain." International journal of medical informatics 62.2 (2001): 103-111. [5] Cheung, David W., et al. "A fast distributed algorithm for mining association rules.". Parallel. and. Distributed. Information. Systems,. 1996.,. Fourth. International Conference on. IEEE, 1996. [6] Chiang, Johannes, and Chia-Chi Wu. "Mining multi-dimensional association rules in multiple database segmentation." Proceedings of International Conference on Information Management. 2005. [7] Chuang, Kun-Ta, Ming-Syan Chen, and Wen-Chieh Yang. "Progressive sampling for association rules based on sampling error estimation." Advances in Knowledge Discovery and Data Mining. Springer Berlin Heidelberg, 2005. 505-515.. 47.

(58) [8] Das, Amitabha, Wee-Keong Ng, and Yew-Kwong Woon. "Rapid association rule mining." Proceedings of the tenth international conference on Information and knowledge management. ACM, 2001. [9] Dean, Jeffrey, and Sanjay Ghemawat. "MapReduce: simplified data processing on large clusters." Communications of the ACM 51.1 (2008): 107-113. [10] Do, Tien Dung, Siu Cheung Hui, and Alvis Fong. "Mining frequent itemsets with. category-based. constraints.". Discovery. Science.. Springer. Berlin. Heidelberg, 2003. [11] Ghemawat, Sanjay, Howard Gobioff, and Shun-Tak Leung. "The Google file system." ACM SIGOPS operating systems review. Vol. 37. No. 5. ACM, 2003. [12] Han, Jiawei, Jian Pei, and Yiwen Yin. "Mining frequent patterns without candidate generation." ACM SIGMOD Record. Vol. 29. No. 2. ACM, 2000. [13] Khare, Neelu, Neeru Adlakha, and K. R. Pardasani. "An Algorithm for Mining Multidimensional Association Rules using Boolean Matrix." Recent Trends in Information, Telecommunication and Computing (ITC), 2010 International Conference on. IEEE, 2010. [14] Lent, Brian, Arun Swami, and Jennifer Widom. "Clustering association rules." Data Engineering, 1997. Proceedings. 13th International Conference on. IEEE, 1997. [15] Li, Lingjuan, and Min Zhang. "The strategy of mining association rule based on cloud computing." Business Computing and Global Informatization (BCGIN), 2011 International Conference on. IEEE, 2011. [16] Li, Ning, et al. "Parallel implementation of apriori algorithm based on MapReduce." Software Engineering, Artificial Intelligence, Networking and. 48.

(59) Parallel & Distributed Computing (SNPD), 2012 13th ACIS International Conference on. IEEE, 2012. [17] Lin, Ming-Yen, Pei-Yu Lee, and Sue-Chen Hsueh. "Apriori-based frequent itemset mining algorithms on MapReduce." Proceedings of the 6th international conference on ubiquitous information management and communication. ACM, 2012. [18] Qiu, Hongjian, et al. "YAFIM: A Parallel Frequent Itemset Mining Algorithm with Spark." Parallel & Distributed Processing Symposium Workshops (IPDPSW), 2014 IEEE International. IEEE, 2014. [19] Schuster, Assaf, and Ran Wolff. "Communication-efficient distributed mining of association rules." Data Mining and Knowledge Discovery 8.2 (2004): 171-196. [20] Srikant, Ramakrishnan, and Rakesh Agrawal. "Mining quantitative association rules in large relational tables." ACM SIGMOD Record. Vol. 25. No. 2. ACM, 1996. [21] Wojciechowski, Marek, and Maciej Zakrzewicz. "Dataset filtering techniques in constraint-based frequent pattern mining." Pattern detection and discovery. Springer Berlin Heidelberg, 2002. 77-91. [22] Yang, Xin Yue, Zhen Liu, and Yan Fu. "MapReduce as a programming model for association rules algorithm on Hadoop." Information Sciences and Interaction Sciences (ICIS), 2010 3rd International Conference on. IEEE, 2010. [23] Zaharia, Matei, et al. "Resilient distributed datasets: A fault-tolerant abstraction for in-memory cluster computing." Proceedings of the 9th USENIX conference on Networked Systems Design and Implementation. USENIX Association, 2012. [24] “Frequent Itemset Mining Dataset Repository”, http://fimi.ua.ac.be/data/ 49.

(60) 50.

(61) ੝‫ݯۓ‬ᕍ໨Ҟжဦ ΋ǵᄌ‫ੰ܄‬Ȑ๏ᛰΎВа΢‫ޣ‬ȑ. 30. Ϻ੺የ. 31. ҜԼ‫ݹ‬. жဦ. Ӝᆀ. 32. ӄ‫܄ي‬ᔸ੶. 01. ᑗֿੰ. 33. ߙӀ౳. 02. ଯՈᓸ. 34. ଳ౳ੱ. 03. ᄌ‫ݹط܄‬. 35. ຎᆛጢᡂ‫܄‬. 04. ᄌ‫܄‬๝᠌‫ݹ‬. 36. ໳ඬ೽ᡂ‫܄‬. 05. Ҙ‫ރ‬ဏᐒૈምᛖ. 37. ဟ๻ጢ‫ݹ‬. 06. ফഹ. 38. ࣒ዟᡏрՈ. 07. ภ॥. 39. ‫ف‬ጢᡂ‫܄‬. 08. ੃ϯ‫܄‬ውᅬ. 40. ᄌ‫܄‬೦Ո. 09. ᜢ࿯‫ݹ‬. 41. ๋ඬੱ. 10. ᄌ‫܄‬Ѝ਻ᆅ‫ݹ‬. 42. η৐ϣጢ౦Տੱ. 11. Ј᠌ੰ. 43. ‫ݜ‬٢નၸଯੱ. 12. ᕎੱᛰ‫ނ‬ଓᙫ‫ݯ‬ᕍ. 44. ϣԸ߻৥ੰᡂ. 13. ᠌Ꮤ౽෌ࡕᛰ‫ނ‬ଓᙫ‫ݯ‬ᕍ. 45. ᄌ‫܄‬ሷᝠ‫ݹ‬. 14. တՈᆅੰᡂ. 46. ᄌ‫܄‬ύԸ‫ݹ‬. 15. ᡜ⻯. 47. ᆒઓੰ. 16. ЃߎහМੱ. 48. ࠶ᅟᇽМੱ. 17. ่ਡੰ. 49. Լሼ‫ޔ‬๺ᕭੱ. 18. ᄌ‫܄‬ᖌၰ‫ݹ‬. 50. ӭว‫܄‬Լ‫ݹ‬. 19. ଯՈિੱ. 51. ख़ੱԼคΚ. 20. ‫਻ޤ‬ဍ. 52. ӃϺ‫܄‬жᖴ౦த੯ੰ. 21. ᄌ‫ߔ܄‬༞‫ݹޤ܄‬. 53. ӃϺǵࡕϺխࣝόӄੱ. 22. Ѝ਻ᆅᘉ஭ੱ. 54. 23. ဉᐒૈ‫܄‬ምᛖȐ֖ᄌ‫܄‬ી᠌ ‫ݹ‬ǵӚᅿगဉ৲Ժੱǵ࡚ចε ဉੱংဂǵगဉᕧឳ‫ੱݹ܄‬ǵ ᄌ‫܄‬εဉ‫ੱݹ‬ȑ. ‫ځ‬дύኰઓ࿶‫س‬಍ᡂ፦Ϸᒪ໺ ‫܄‬੯ੰ. 55. ӭว‫܄‬ฯϯੱ. 56. ᓻ‫ٽ‬တ‫܄‬ഞ࿄Ϸ‫ځ‬дഞ࿄‫܄‬᝛ ংဂ. 24. आඬ‫܄‬ਫ਼የ. 57. ୏ેๆኬฯϯ. 25. ‫ط‬ฯϯ. 58. ୏ેਵ༞ϷՈਵੱ. 26. ႜᒍМੱ. 59. ‫ޤ‬؈๱ੱ. 27. ମ፦౧᚞ੱ. 60. ѦӢ‫܌‬ठϐ‫ޤ‬੯ੰ. 28. ଳ᠆. 61. ๝᠌གࢉ. 29. Ҙ᠆. 62. ӃϺ‫܄‬྾‫׎‬੯ੰ. B-56.

(62) 63. ࡭ុ‫܄‬ՈనᏉ‫ڰ‬ምᛖȐՈ϶ੰȑ. 95. ᄌ‫܄‬ମᡎ‫(ݹ‬87/4/1). 64. ࿂ᅭ. 96. ମᡎϩϯό‫ੱؼ‬ংဂ(88/4/1). 65. ౨የ. চว‫܄‬Ոλ݈ቚғੱ(88/4/1). 66. ឪៈဏȐ߻ӈဏȑ‫ޥ‬ε. 97 ຏǺ. 67. ଶ࿶ੱংဂ. 68. ֿѨ࿣. 69. ‫ੱݨ‬Ȑӭෛᖄशύࢥȑ. 70. ๝΢ጕੰᡂЇวϣϩ‫ݜ‬ምᛖ. 71. တΠࠟᡏੰᡂЇวϣϩ‫ݜ‬ምᛖ. 72. ‫܄‬Ԑዕ. 73. တዦ‫ٳ‬วઓ࿶фૈምᛖ. 74. ӭว‫ڬ܄‬ᜐઓ࿶ੰᡂ. 75. ઓ࿶ᘀੰᡂ. A1. ຬॣ‫ݢ‬ᔠࢗ. 76. Οΰઓ࿶ภ. A2. Ըሷോࣽᔠࢗ. 77. ୃᓐภ. A3. ຎ᜔ᔠࢗ. 78. ο஘ੰ‫ٳ‬วЈ᠌Ոᆅ౦த‫ޣ‬. A4. ੰ౛ಔᙃᔠࢗ. 79. ਜ਼ဌੰ. A5. ਡηᙴᏢᔠࢗ. 80. ୋҘ‫ރ‬ဏᐒૈեΠੱ. A6. ̔Ӏᔠࢗ. 81. ૉᡎཞ໾. A7. ੝ਸ೷ቹᔠࢗ. 82. ၸ௵‫܄‬ሷ‫ݹ‬. A8. ઓ࿶ࣽᔠࢗ. 83. қඬ. 84. િᅅ‫܄‬Ҝጥ‫ݹ‬. 85. ᜪᐘણ؈ᑈੱȐज़ੰ‫؞‬ຬၸᡏ ߄य़ᑈԭϩϐΟΜа΢‫ޣ‬ȑ. Οǵ੝ਸ‫ݯ‬ᕍ‫܈‬ೀ࿼Ǻ. 86. ᜪϺ੺የ(86/1/1). D1. ᕎੱ‫ܫ‬৔ጕ‫ݯ‬ᕍ. 87. ੺੶‫܄‬Ҝጥ‫(ݹ‬86/1/1). D2. ᕎੱϯᏢ‫ݯ‬ᕍ. 88. ৎ௼‫܄ؼ܄‬ᄌ‫܄‬Ϻ੺የ(86/1/1). D3. 89. ߄Ҝϩှ‫܄‬Н੺ੱ(86/1/1). 90. ᝄख़‫܄‬ങᡒ᠆Ȑ֖ቫ‫ރ‬ങᡒ᠆. D4. ᆒઓࣽ‫ݯ‬ᕍ. Ϸങᡒ᠆‫ރ‬आҜੱȑ(86/1/1). D5. ଯᓸ਼‫ݯ‬ᕍ. 91. Л៶‫ف‬ϯੱ(86/1/1). D6. ౳្ࣽ৔‫ݯ‬ᕍ. 92. ຾Չ‫܄‬ӄ‫ي‬ฯҜੱ(86/1/1). D7. Ո϶ੰ‫ݯ‬ᕍ. 93. ‫܄‬ဏեૈੱȐHypognadismȑ. D8. Ոన೸‫ݯ݋‬ᕍ. D9. ဎጢ೸‫݋‬. 94. (86/7/1) ᄌ‫܄‬ឪៈဏ‫ݹ‬Ȑሡ࿶ឪៈဏࡪ. D0. ‫ނ‬౛‫ݯ‬ᕍᙁൂǵύࡋ‫ݯ‬ᕍ. ύᙴᄌ‫(ੰ܄‬๏ᛰΎВа΢‫)ޣ‬ ੝‫ݯۓ‬ᕍ໨Ҟжဦࣁ΢ॊጓዸ 01 Կ 12ǵ14ǵ19ǵ25ǵ27ǵ32ǵ 45ǵ65ǵ67ǵ69ǵ76ǵ77 ฻ 23 ໨. Βǵ੝ਸᔠࢗǺ. ൺ଼‫ݯ‬ᕍȐ‫ނ‬౛‫ݯ‬ᕍᙁൂǵύ ࡋ‫ݯ‬ᕍନѦȑ. Ȑ88/8ȑ. ነ‫ڗ‬ឪៈဏϩ‫ݜ‬న᛾ჴ‫ޣ‬ȑ (86/7/1) B-57.

(63) Уᙴᙴᕍ୍ܺ):9/2 ቚु*. ѤǵУᙴǺ P1. ਥᆅ‫ݯ‬ᕍ. P2. ሌણк༤. P3. ፄӝᐋિȐ࣒ዟዟηȑк༤. P4. FG. ଣ‫୍ཱུܺ܌‬ख़ࡋߚᆒઓምᛖ ‫ޣ‬Уᙴᙴᕍ୍ܺ!. FH. ଣ‫୍ܺ܌‬ख़ࡋߚᆒઓምᛖ‫ޣ‬ Уᙴᙴᕍ୍ܺ!. У‫ੰڬ‬ЋೌȐ֖Ꮑ៱Π‫ڊ‬ନ ೌȑ. FI. ଣ‫୍ܺ܌‬ύࡋߚᆒઓምᛖ‫ޣ‬. P5. ‫ูٽ‬ᘐᡎೀ౛. FJ. P6. ଯᓸ਼‫ݯ‬ᕍ. ଣ‫୍ܺ܌‬ᇸࡋߚᆒઓምᛖ‫ޣ‬ Уᙴᙴᕍ୍ܺ!. P7. α๚ѦࣽߐບЋೌȐхࡴ‫ܘ‬. FK. ᙴᕍი୍ཱུܺख़ࡋߚᆒઓም ᛖ‫ޣ‬Уᙴᙴᕍ୍ܺ!. FL. ᙴᕍი୍ܺख़ࡋߚᆒઓምᛖ ‫ޣ‬Уᙴᙴᕍ୍ܺ!. Уᙴᙴᕍ୍ܺ!. Уȑ P8. ‫ݯ‬ᕍ‫܄‬У่ҡమନ. F2. УᙴԿคУᙴໂ୺཰ीฝ. F3. УᙴৣคУᙴໂْ଑ᙴᕍܺ ୍. FM. ᙴᕍი୍ܺύࡋߚᆒઓምᛖ ‫ޣ‬Уᙴᙴᕍ୍ܺ!. F4. ӃϺ‫܄‬যᚓ຋஻‫ޣ‬Уᙴᙴᕍ. FN. ᙴᕍი୍ܺᇸࡋߚᆒઓምᛖ ‫ޣ‬Уᙴᙴᕍ୍ܺ!. ୍ܺȐ91/1ȑ F5 F6. ख़ࡋа΢‫ي‬Јምᛖ‫ޣ‬Уᙴᙴ. FS. ᕍ୍ܺȐ91/1ȑ Ȑ99/01/01 ‫੃ڗ‬ȑ. ੝ਸᙴᕍ୍ܺ၂ᒤीฝϐУ. ΟྃȐ֖ȑаΠᓻѴ‫ٽ‬᡹Ꮑٛ ‫(୍ܺڋ‬91/1)Ȑ94/1 ‫੃ڗ‬ȑ. ᙴ‫ډ‬ӻᙴᕍ୍ܺ!. F7. ଣ‫܌‬ύࡋ‫ي‬Јምᛖ‫ޣ‬Уᙴᙴ ᕍ୍ܺȐ99/01/01 ‫੃ڗ‬ȑ. F8. ᙴᕍი໗ख़ࡋа΢‫ي‬Јምᛖ ‫ޣ‬Уᙴᙴᕍ୍ܺȐ99/01/01 ‫ڗ‬ ੃ȑ. F9. ᙴᕍიύࡋ‫ي‬Јምᛖ‫ޣ‬Уᙴ. ଣ‫ݯ܌‬ᕍख़ࡋᆒઓምᛖ‫ޣ‬У ᙴᙴᕍ୍ܺ):9/2 ቚु*. FD. ଣ‫ݯ܌‬ᕍύࡋᆒઓምᛖ‫ޣ‬У ᙴᕍი‫ݯ‬ᕍख़ࡋᆒઓምᛖ‫ޣ‬ Уᙴᙴᕍ୍ܺ):9/2 ቚु*!. FF. FU. ӄ଼҇நߥᓀУᙴߐບᕴᚐ ੝ਸᙴᕍ୍ܺ၂ᒤीฝϐ੝ ‫ۓ‬ሡ‫)୍ܺޣ؃‬213/12/12 ཥ жဦࣁ F2ǵF3ǵFAǵFT ‫ޣ‬Ǵਢ ҹϩᜪࣁ 14ǹжဦࣁF4ǵF5ǵF6ǵ F7ǵF8ǵF9ǵFCǵFDǵFEǵFF ‫ޣ‬Ǵӄ଼҇நߥᓀУᙴߐບᕴᚐ ੝ਸᙴᕍ୍ܺ၂ᒤीฝϐଣ‫ܺ܌‬ ୍ϐଣ‫୍ܺ܌‬Ǻཱུख़ࡋ FGǵख़ࡋ FHǵύࡋ FIǵᇸࡋ FJǴᙴᕍი ୍ܺǺཱུख़ࡋ FKǵख़ࡋ FLǵύ ࡋ FMǵᇸࡋ FN(99.1 ቚु)FSǵ. ᙴᙴᕍ୍ܺ):9/2 ቚु* FE. УᙴৣԿУᙴᙴᕍၗྍό‫ى‬ Ӧ୔ْ଑୍ܺीฝ.‫୔ޗ‬ᙴᕍ ઠȐ212/12 ཥቚȑ!. !. ଭ઒Ӧ୔У‫ྣੰڬ‬ៈᆛ၂ᒤ ीฝ(97/1) Ȑ100/01 ‫੃ڗ‬ȑ. FC. FT. ቚ*!. ᙴᕍ୍ܺȐ99/01/01 ‫੃ڗ‬ȑ FA. ӄ଼҇நߥᓀУᙴߐບᕴᚐ. FU(102.1)Ǵਢҹϩᜪࣁ 16ǶFP У ‫ੰڬ‬಍ӝྣៈಃ΋໘ࢤǵFQ У‫ੰڬ‬಍ӝ. ᙴᕍი‫ݯ‬ᕍύࡋᆒઓምᛖ‫ޣ‬ B-58.

(64) ྣៈಃΒ໘ࢤǵFR У‫ੰڬ‬಍ӝྣៈಃΟ. ДҔᛰໆਢҹȐ212/22 ཥ. ໘ࢤǴਢҹϩᜪࣁ15(ଛӝ 99.01.12 ଼ߥᙴӷಃ 0990071960 ဦϦ֋ཥ. ቚȑ Ƕ CE. ቚ)Ƕ. ύᙴ.࡭ᄌ‫ੰ܄‬ೱុೀБᆇሦ ᛰǴςрੇࣁᇻࢩᅕಭբ཰ಭ ঩Ǵග‫ٮ‬Ϫ่ЎҹǴ΋ԛሦ‫ڗ‬ 3 ঁД‫ ܈‬4 ঁДҔᛰໆਢҹ Ȑ212/22 ཥቚȑ Ƕ. ϖǵύᙴ੝ਸ‫ݯ‬ᕍ‫܈‬ೀ࿼Ǻ. CF. ύᙴ.࡭ᄌ‫ੰ܄‬ೱុೀБᆇሦ. C1. ύ॥ࡕᒪੱ. ᛰǴςрੇࣁ୯ሞૐጕಭಬբ. C2. ᄌ‫܄‬ሷ‫ݹ‬. C3. ଞ‫ء‬. ཰ಭ঩Ǵග‫ٮ‬Ϫ่ЎҹǴ΋ԛ ሦ‫ ڗ‬3 ঁД‫ ܈‬4 ঁДҔᛰໆਢ. C4. ໾ࣽ‫ݯ‬ᕍ. ҹȐ212/22 ཥቚȑ Ƕ. C5. ಥՀ᏾ൺ. C6. ύᙴᙴᕍၗྍό‫ى‬Ӧ୔ْ଑. CG. ЎҹǴ΋ԛሦ‫ ڗ‬3 ঁД‫ ܈‬4 ঁ ДҔᛰໆਢҹȐ212/22 ཥ ቚȑ Ƕ. ᙴᕍीฝȐচӜǺคύᙴໂْ ଑ᙴᕍ ȑ C7. ύᙴ.࡭ᄌ‫ੰ܄‬ೱុೀБᆇሦ ᛰǴ‫ـش‬੯ੰੰΓǴග‫ٮ‬Ϫ่. ύᙴᙴᕍၗྍό‫ى‬Ӧ୔ዛᓰ ໒཰ीฝȐচӜǺคύᙴໂዛ. J1. ύᙴ.Չ୏όߡ‫ޣ‬Ǵ࿶ᙴৣᇡ‫ۓ‬. ᓰ໒཰ȑ. ‫܈‬࿶‫ૼڙ‬Γග‫ٮ‬Ϫ่ЎҹǴᄌ. C8. တՈᆅ੯ੰϐՋᙴՐଣੰ஻ ύᙴᇶշᙴᕍȐ98 ԃଆ‫੃ڗ‬ȑ. ‫ੰ܄‬жሦᛰਢҹ):7/8 ቚुǹ 212/22 Ўӷঅु*. J8. ᒀየϐՋᙴՐଣੰ஻ύᙴᇶ շᙴᕍ. C9. λ‫਻ٽ‬ഹ጗ှයύᙴᓬ፦ߐ ບྣៈ၂ᒤीฝ. C0. λ‫ٽ‬တ‫܄‬ഞ࿄ύᙴᓬ፦ߐບ. J2. ύᙴ.ςрੇࣁᇻࢩᅕಭբ཰ ಭ঩Ǵග‫ٮ‬Ϫ่ЎҹǴᄌ‫ੰ܄‬ жሦᛰਢҹ):7/8 ቚुǹ 212/22 Ўӷঅु*. J3. ύᙴ.ςрੇࣁ୯ሞૐጕಭಬ բ཰ಭ঩Ǵග‫ٮ‬Ϫ่ЎҹǴᄌ. ྣៈ၂ᒤीฝ CA. တՈᆅࡕᒪੱ)::/2 ཥቚ*. CB. ႴᓰୃᇻӦ୔ύᙴৣߏය຾ Ꭻ၂ᒤीฝ. CC. ‫ੰ܄‬жሦᛰਢҹ):8/21 ቚ ुǴ212/22 Ўӷঅु* J4. ‫׎‬Ǵᄌ‫ੰ܄‬жሦᛰਢҹ Ȑ212/22 ཥቚȑ Ƕ. ύᙴ.࡭ᄌ‫ੰ܄‬ೱុೀБᆇሦ ᛰǴႣ‫ۓ‬р୯Ǵග‫ٮ‬Ϫ่Ў ҹǴ΋ԛሦ‫ ڗ‬3 ঁД‫ ܈‬4 ঁД. CD. ύᙴ.࿶ߥᓀΓᇡ‫ۓ‬ϐ੝ਸ௃. ҔᛰໆਢҹȐ212/22 ཥቚȑ Ƕ. ဍዦ஻‫ޣ‬Ћೌǵϯᕍǵ‫ܫ‬৔ጕ ᕍ‫ࡕݤ‬ՋᙴՐଣύᙴᇶշᙴ. ύᙴ.࡭ᄌ‫ੰ܄‬ೱុೀБᆇሦ ᛰǴ߇ӣᚆ৞Ӧ୔Ǵග‫ٮ‬Ϫ่. !. J7. ᕍ၂ᒤीฝ):9/12 ቚु*Ƕ. ЎҹǴ΋ԛሦ‫ ڗ‬3 ঁД‫ ܈‬4 ঁ. Ϥǵ‫ځ‬д B-59.

(65) E1. ဉੰࢥ(87). E2. ᙴৣЍජԿӼᎦǵᎦៈǵූ ምǵᅽճᐒᄬϷៈ౛ϐৎ࣮ບ ਢҹȐ89/3/16ȑ. G7. Бਢ—ς໒཰Ȑ93/01ȑ. Ѝජߏයྣៈᐒᄬග‫ٮ‬΋૓ߐ ບਢҹ(99.1 অु) E3 E4. ᙴᕍၗྍલЮӦ୔୍ܺዛᓰ. G8. ৎ৥ᙴৣ᏾ӝ‫ྣ܄‬ៈ Ȑ92/3/10ȑ. G9. ξӦᚆ৞Ӧ୔ᙴᕍ๏бਏ੻ ගܹीฝȐ93/01ȑ. ่ਡੰ၂ᒤीฝ Ȑ93/01~97/01ȑ. H1. BǵC ࠠ‫ݹط‬၂ᒤीฝ(93/01). H2. ᑗֿੰ၂ᒤीฝǺаӅӕྣៈ. Ջᙴ.Չ୏όߡ‫ޣ‬Ǵ࿶ᙴৣᇡ‫܈ۓ‬ ࿶‫ૼڙ‬Γග‫ٮ‬Ϫ่ЎҹǴᄌ‫܄‬. ኳԄࣁ୷ᘵ‫ޑ‬ᑗֿੰΓ੯ੰ ᆅ౛ჴᡍीฝ (ч୔ϩֽ. ੰжሦᛰਢҹ):7/8 ቚुǹ 212/22 অुЎӷ*. 89/5/17 ၂ᒤ) Ȑ90/11 ӄय़၂ᒤȑ. H3. Ջᙴ.ςрੇࣁᇻࢩᅕಭբ཰ ಭ঩Ǵග‫ٮ‬Ϫ่ЎҹǴᄌ‫ੰ܄‬ жሦᛰਢҹ):7/8 ቚुǺ 212/22 অुЎӷ*. E5. ࠄ୔ϩֽǺຼౢය߻ᘳ‫܄‬Ѝб ‫ࡋڋ‬၂ᒤीฝ(90/9). E6. ਻ഹ၂ᒤीฝȐ90/11ȑ. E7. ‫่ޤ‬ਡ၂ᒤीฝ Ȑ90/11~93/09ȑ. H4. Ծ຤଼ᔠว౜ੰӂуբೀ࿼ ‫܈‬ᔠࢗ(97/1). E8. ଯՈᓸ၂ᒤीฝȐ95/1ȑ. H5. жᖴੱংဂྣៈБਢȐ97.8 ቚ. E9. ਜ਼ဌੰ၂ᒤीฝȐ95/10 ‫੃ڗ‬ȑ. EA. Ѝජߏයྣៈᐒᄬග‫ٮ‬ൺ଼ ‫ݯ‬ᕍਢҹ(99.1 ቚु). EB. ӄ଼҇நߥᓀ߃යᄌ‫܄‬๝᠌ ੰᙴᕍ๏б‫ׯ‬๓ीฝ(100.01 ቚु). F1. ुȑ H6. Ջᙴ.ςрੇࣁ୯ሞૐጕಭಬ բ཰ಭ঩Ǵග‫ٮ‬Ϫ่ЎҹǴᄌ ‫ੰ܄‬жሦᛰਢҹ):8/21 ቚ ुǹ212/22 অुЎӷ*. H7. ӄ଼҇நߥᓀ B ࠠ‫ݹط‬஥চ ‫ޣ‬Ϸ C ࠠ‫ݹط‬གࢉ‫ޣ‬ᙴᕍ๏ б‫ׯ‬๓Бਢ(99.1 ቚु). ࠄ‫׫‬ᑜߞကໂᄤϘངໂᆒઓ ੯ੰᙴᕍ૶ॅਏ੻ගϲीฝ H8. Ջᙴ.࡭ᄌ‫ੰ܄‬ೱុೀБᆇሦ. F6. (90/8) ΟྃȐ֖ȑаΠᓻ‫ٽ‬᡹Ꮑٛ‫ڋ‬ ୍ܺ(91/1)Ȑ94/1 ‫੃ڗ‬ȑ. F7. Ңጄߐບ၂ᒤीฝ(91/1). ᛰǴႣ‫ۓ‬р୯Ǵග‫ٮ‬Ϫ่Ў ҹǴ΋ԛሦ‫ ڗ‬3 ঁД‫ ܈‬4 ঁД ҔᛰໆਢҹȐ212/22 Ўӷঅ. Ȑ93/10 ‫੃ڗ‬ȑ. ुȑ Ƕ. G4. ፁғᅽճ೽ᆙ࡚ᙴᕍၗྍલЮӦ୔ ‫ׯ‬๓ीฝȐ95/04ȑ. G5. ᙴᕍၗྍલЮӦ୔୍ܺዛᓰ Бਢ—ْ଑ᙴᕍȐ93/01ȑ. G6. ᙴᕍၗྍલЮӦ୔୍ܺዛᓰ. H9. Ջᙴ.࿶ߥᓀΓᇡ‫ۓ‬ϐ੝ਸ௃ ‫׎‬Ǵᄌ‫ੰ܄‬жሦᛰਢҹ Ȑ212/22 ཥቚȑ Ƕ. HA. Ջᙴ.࡭ᄌ‫ੰ܄‬ೱុೀБᆇሦ ᛰǴ߇ӣᚆ৞Ӧ୔Ǵග‫ٮ‬Ϫ่ ЎҹǴ΋ԛሦ‫ ڗ‬3 ঁД‫ ܈‬4 ঁ ДҔᛰໆਢҹȐ212/22 ཥ. Бਢ—ཥ໒཰Ȑ93/01ȑ B-60.

(66) ቚȑǶ HB. Ջᙴ.࡭ᄌ‫ੰ܄‬ೱុೀБᆇሦ ᛰǴςрੇࣁᇻࢩᅕಭբ཰ಭ ঩Ǵග‫ٮ‬Ϫ่ЎҹǴ΋ԛሦ‫ڗ‬ 3 ঁД‫ ܈‬4 ঁДҔᛰໆਢҹ Ȑ212/22 ཥቚȑ Ƕ. HC. Ջᙴ.࡭ᄌ‫ੰ܄‬ೱុೀБᆇሦ ᛰǴςрੇࣁ୯ሞૐጕಭಬբ ཰ಭ঩Ǵග‫ٮ‬Ϫ่ЎҹǴ΋ԛ ሦ‫ ڗ‬3 ঁД‫ ܈‬4 ঁДҔᛰਢҹ Ȑ212/22 ཥቚȑ Ƕ. HD. Ջᙴ.࡭ᄌ‫ੰ܄‬ೱុೀБᆇሦ ᛰǴ‫ـش‬੯ੰੰΓǴග‫ٮ‬Ϫ่ ЎҹǴ΋ԛሦ‫ ڗ‬3 ঁД‫ ܈‬4 ঁ ДҔᛰਢҹȐ212/22 ཥቚȑ Ƕ. JA. ԏ৒ჹຝᙴᕍ୍ܺीฝ.ᕖ҅ ᐒᜢϣߐບȐ213/2 ଆ፾Ҕȑ. JB. ԏ৒ჹຝᙴᕍ୍ܺीฝ.‫י‬ៈ ߐບ)213/2 ଆ፾Ҕ*. K1. ӄ଼҇நߥᓀ Qsf.FTSE Ⴃٛ ‫܄‬ीฝϷੰΓፁ௲ीฝ )212/2 ቚु*Ƕ. N. ٢ᕎ၂ᒤीฝཥঁਢ(90/11). C. ٢ᕎ၂ᒤीฝൺวঁਢ(90/11). R. ٢ᕎ၂ᒤीฝֹԋঁਢ(90/11). ! жဦࣁ E3ǵE4ǵE5ǵE6ǵE7ǵE8ǵ E9ǵF1ǵNǵCǵR ǵH1ǵH5ǵH7 ‫ޣ‬Ǵ ਢҹϩᜪࣁE1ǹжဦࣁG5ǵG6ǵG7 ‫ޣ‬Ǵ ਢҹϩᜪࣁ D4ǹжဦࣁ G4ǵG8ǵ G9 ‫ޣ‬Ǵ٩‫܌‬ឦϐਢҹϩᜪҙൔǶ. B-61.

(67)