Publish to my blog (weekly)

• 4.3. Aggregating functions - Chapter 4. Functions

  • percentileCont(n.age, 0.4)
    • PERCENTILE_CONT는 순서가 지정된 값 사이의 선형 보간을 계산합니다. 이 함수는 집계 그룹에서 백분위 값(P)과 NULL을 제외한 행들의 번호(N)를 사용하여 정렬 명세에 따라 행의 순서를 지정한 후 행 번호를 계산합니다. 행 번호(RN)를 계산하는 공식은 RN = (1+ (P*(N-1))입니다. 이 집계 함수의 최종 결과는 행 번호가 CRN = CEILING(RN)과 FRN = FLOOR(RN)인 행의 값 사이 선형 보간을 통해 계산됩니다.최종 결과는 다음과 같습니다.(CRN = FRN = RN)일 때 결과는 (value of expression from row at RN)입니다.그렇지 않다면 다음 결과가 표시됩니다.(CRN - RN) * (value of expression for row at FRN) + (RN - FRN) * (value of expression for row at CRN).> var DATA = [13, 33, 44]> var N = 3> var P = 0.4> var RN = 1 + (P * (N - 1))> RN1.8> var CRN = Math.ceil(RN)> CRN2> var FRN = Math.floor(RN)> FRN1> var PC = (CRN - RN) * DATA[FRN - 1] + (RN - FRN) * DATA[CRN - 1]> PC29
  • percentileDisc(n.age, 0.5)
    • PERCENTILE_DISC는 이산 분포 모델을 가정하는 역분포 함수로서 백분위 값과 정렬 명세를 가지며, 지정된 집합에서 요소를 반환합니다.임의의 백분위 값을 P라고 할 때, PERCENTILE_DISC는 ORDER BY 절의 표현식 값을 정렬한 후 동일한 정렬 명세와 관련하여 가장 작지만 P보다는 크거나 같은 누적 분포 값을 반환합니다.

• 趙相學

  趙相學

• Graph Databases for Beginners: Data Modeling Pitfalls to Avoid

  • However, this data model isn’t robust enough to detect wrongful behavior.
  • Even with properties attached to each EMAILED relationship, we wouldn’t be able to correlate connections between EMAILED, CC and BCC relationships – and those correlating relationships are exactly what we need for our fraud detection solution.

• What Will the Future of RPA Look Like?

  • RPA technologies are positioned to become even more integrated with cloud services in the future.
  • While the automation enabled by RPA can, of course, still be delivered as an onsite software package, moving business processes to remote servers as part of the cloud will allow many businesses to obtain greater operational productivity and efficiency than has ever been possible in the past.
  • RPA robots will be able to improve their own performance and make decisions through the incorporation of intelligent technologies like machine learning.
  • Through machine learning algorithms, computer systems are able to improve their performance by incorporating Big Data and adapting to new rules based on previous experiences without human intervention.
  • unforeseen errors
  • detected data patterns
  • exceptions in business processes
  • scenarios never before seen

Posted from Diigo. The rest of my favorite links are here.