Cluster Analysis On The Mobile Phone Brand Series Based On Principal Component

The widespread use of mobile phones as a tool of communication has brought usgreat convenience in work, study and entertainment. However, facing so manymobile phone brands and series, it is easy falling into troubles for a consumer inchoosing a product and it is also difficult for a producer is in grasping consumers’preferences. A "violent war" in saling markets brings both the consumers andproducers with great pressure, making the cell phone market much more challenging.In this paper, We have collected data consisting of16factors, which consumers arecommonly concerned about, such as screen size, weight, pixel number of camera andso on, for about50series of8main brands of mobile phones at home and abroad.Using SPSS tools for data mining, we extracts the principal component of theconcerned factors, analyzes the mobile phone brand series based on the principalcomponent factors, and constructs a hierarchical clustering chart, in order to providesome reference for both consumers and producers.This paper firstly summarizes the limitations of previous research work onmobile phone brand clustering. It then chooses the mobile phone features that arehighly related with its price to defined and quantificated. Considering that collecteddata is generally incomplete, we use the multiple stochastic interpolation method tomake up the missing data. We then find the principal components of linearindependence by the factor analysis dimension reduction technique, and compute themobile phone series scores according to the main components. The hierarchicalclustering is subsequently performed based on the scores so that we can rule out theproblem that linear correlation parameters of mobile phone can’t be clustered. Finally,according to the clustering results and combining with the mean values of parametersof all kinds of mobile phone brands, we divide the market into six characteristicsegments.In summary, based on analyzing the mobile phone clustering methods, the authorputs forward his ideas and views, and formulates them into practical analysisapproach, hoping that the new method can make some contribution to the actual market segmentation of cell phones.