[Hair] Cluster Analysis

너무 내러티브함

청중에게 question을 던지는 형식 필요함

발표시간 조절

발음 공부 열심히 - 천천히 해도 괜찮으 

 

What is cluster analysis?

purpose: group objects based on the characteristics

referres as

Q analysis, typology construction, classification analysis, and numerical taxonomy

why?

the usage of clustering methods in such diverse disciplines

 

VS factor analysis

 makes the groupings based on patterns of variation (correlation)

<-> whereas cluster analysis makes groupings on the basis of distance (proximity).

 

Cluster Analysis as a Multivariate Technique

classifies objects

ex) respondents, products ..

 

high internal (within-cluster) homogeneity

high external (between-cluster) heterogeneity

 

Cluster analysis

: the only multivariate technique that does not estimate the variate empirically

<-> uses the variate as specified by the researcher.

 

focus

: the comparison of objects based on the variate

 

Conceptual Development with Cluster Analysis

common roles cluster analysis

 

1. Data reduction

a large number of meaningless observation

-> should reduce the information from an entire population or sample

 

2. Hypothesis generation

useful when a researcher wishes to develop hypotheses concerning the nature of the data or to examine previously stated hypotheses.

ex)

attitudes toward the consumption of diet VS regular soft drinks

separate soft-drink consumers into logical segments

-> classify soft-drink consumers by their attitudes about diet VS regular soft drinks

=> profiled for demographic similarities and differences.

 

Necessity of Conceptual Support in Cluster Analysis

cluster analysis can be criticized for working too well in the sense that statistical results are produced even when a logical basis for clusters is not apparent

-> researcher should have a strong conceptual basis

ex) why groups exist in the first place

what variables logically explain why objects end up in the groups that they do

 

most common criticisms that must be addressed by conceptual rather than empirical support

1. descriptive, atheoretical, and noninferential

2. always create clusters, regardless of the actual existence of any structure in the data

: the researcher is making an assumption of some structure among the objects. 

-> The researcher should always remember that just because clusters can be found does not validate their existence.

 

3. not generalizable

resaon: dependent upon the variables used as the basis for the similarity measure.

 

How does cluster analysis work?

- measuring similarity, forming clusters, and deciding on the number of clusters that best represent structure. W

- balance of objective and subjective considerations that must be addressed by any researcher

 

Simple Example

The primary objective of cluster analysis is to define the structure of the data by placing the most similar observations into groups.

To accomplish this task, we must address three basic questions:

1. How do we measure similarity? - Measuring similarity

comparing observations on the two clustering variables

: correlation

: distance between observations indicates similarity.

 

2. How do we form clusters? - Forming clusters

the procedure must group those observations that are most similar into a cluster, thereby determining the cluster group membership of each observation for each set of clusters formed

 

3. How many groups do we form? 

trade-off:

fewer clusters and less homogeneity within clusters VS larger number of clusters and more within-group homogeneity

-> a balance must be made between defining the most basic structure (fewer clusters) that still achieves an acceptable level of heterogeneity between the clusters.

 

marketing researcher determine market segments in community based on patterns of loyalty to brands and stores

pilot test

: small sample of seven respondents

1. Measuring similarity

 

 

 

 

 

 

 

Term

Heterogeneity(이질성)"

다양성이나 다양성의 정도를 나타내는 용어입니다. 특히, 통계적이나 연구적인 맥락에서 사용될 때, 이질성은 연구 대상 집단 또는 샘플 내의 다양한 속성, 특성, 또는 행동 양식들의 존재를 나타냅니다. 이질성이 높은 집단이나 샘플은 서로 다른 개인들이나 개체들 간에 큰 차이가 있음을 의미하며, 이는 연구나 분석을 진행할 때 고려해야 할 중요한 요소 중 하나입니다.

 

Aggregate(집합체)"

복수의 개체 또는 요소들이 모여서 이루는 전체를 나타냅니다. 이는 데이터 분석이나 클러스터링 알고리즘에서 주로 사용되며, 여러 개의 개별 항목이나 요소들이 통합되어 단일 단위로 묶이는 것을 의미

 

Pilot test

새로운 제품, 서비스, 또는 기술을 소량의 사용자 집단에게 처음으로 도입하기 전에 시험하는 것을 말합니다. 이는 제품이나 서비스가 실제로 출시되기 전에 사용자들의 피드백을 수집하고 제품 또는 서비스를 개선하기 위한 기회를 제공합니다.

 

Voca

homogeneous: 동질의

heterogeneity: 이질성, 이류

proximity: 가까움, 근접

in the fashion: 이런 방식으로

demographic: 인구 통계학의

profiled for: 개요가 만들어지다, 윤곽이 생기다?

suprious: 그럴싸한, 거짓의

deletion: 삭제

subjective: 주관적인

strive toward: 얻으려고 하다

descriptively: 기술적으로, 묘사적으로

 

---

Hair chapter 3. Exploratory Factor Analysis -

 

처음부터 연구에서 2분법으로 갈라짐

 

EFA

데이터가 기준!

데이터 관점으로 설명!

R이 중요한 것이 데이터가 내가 주장하는 것을 얼마나 설명하고 있느냐

일반적인 다변량 분석은 EFA임!

모르는 용어가 나올시 당연히 검색해서 알아봐야 함

 

CFA

가설을 기준 

기준으로 잡으려면 Theory 기반임

data driven은 할 수 없음 ~!~!

데이터를 부어서 

내 데이터가 가설에 FIT하는지 봄

AMOS가 툴 대표적인 TOOL임

가설에 FIT한 데이터를 찾는 것

 

너무 설명적인 느낌

발표를 위한 발표보다는 남에게 설명하기

발표 좀 더 빨리하기