NumericalDynamics.Com is an organization that provides innovative solutions.

Instruction 6

<<Previous, 1, 2, 3, 4, 5, 6, 7, 8, 9, Next >>

Example 2 - Working with Appliance Sales
Sales managers are always required to understand consumerfs needs, marketable products or regional purchasing properties, and hence need to deal with enormous amount of sales data. PCA will help them interpret such big data easily. In this tutorial, let's analyze the regional purchasing properties using PCA, PLS-DA and PLS-EDA.

First open Appliance Sales Data File. Double click and open Sample2.xlsx.

This data set includes appliance sales of the states of USA. This data is prepared for this tutorial and is completely fictitious. Do not use this data outside of this training. Appliances (variables) are arranged in the first column, cities (samples) are arranged in the second row and states (categories) are shown in the first row.

Data Normalization
First, click Add-Ins tab (if it is not selected), then click gMultibase_2015h and gOpen Formh to open Preparation dialog box.

When Preparation dialog box appears, you set variable range (drag Blue-ray Players to Electric clocks, and click gSet Variable Rangeh. Range g$A$3:$A$65h will be set in variable box.). Likewise set sample range (drag Los Angeles to Vancouver, and click gSet Sample Rangeh. Range g$B$2:$U$2h will be set at sample box). Finally drag California to Washington and click "Set Category Range", then "$B$1:$U$1"will be set at category box as shown below. Although we categorized samples at "Samples" tab in previous tutorial, , we can do it using "Set Category Range" if the categories are already on the sheet.

Next, click "Preparation" tab and choose gData is scaled by observed values" and "Average of all samples is shifted to zero". You can see the reason why we choose them if you open "Help" in this dialog box.

Then click "Next>>" to generate new sheet "Multibase_Preparation".

Basic statistical data, maximum, minimum, average, standard deviation and coefficient of variance (CV) of raw data (before normalizing) are shown in the right side of normalized data.

<<Previous, 1, 2, 3, 4, 5, 6, 7, 8, 9, Next >>



2-3-1 Marunouchi Chiyoda Tokyo Japan