Method of multi-dimensional analysis and display for a large volume of record information items and a system therefor

• US 5,179,643 A
• Filed: 12/07/1989
• Issued: 01/12/1993
• Est. Priority Date: 12/23/1988
• Status: Expired due to Term

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Mori, Hiroshi, Aizawa, Takayuki, Hirata, Shigeki, Homma, Koichi, Kagami, Akira, Akashi, Kichizo
• Primary Examiner(s): Harkcom, Gary V.
• Assistant Examiner(s): Zimmerman, Mark K.
• Application Number:
  US07/447,374
• Time in Patent Office: 1,132 Days
• Field of Search: 364/401-403, 364/518, 364/521, 364/147, 395/140, 395/141, 395/160
• US Class Current: 345/440
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06T 11/00 2D [Two Dimensional] image generation

Abstract

Method and system for extracting a characteristic from information handled by a computer and displaying the information in a manner to clarify the characteristic. A relationship among information items is analyzed from similarity and cluster analysis, and a characteristic related to the relationship is analyzed and displayed. The present invention is particularly useful in the analysis of merchandise information.

Claims

1. A method for displaying a plurality of information items including display data and attributes thereof, comprising steps of:

extracting, from said plurality of information items, first information items contained in said plurality of information items and second information items having a high similarity with respect to an attribute of said first information items, for each of said plurality of information items;

preparing a directed graph having nodes corresponding to said plurality of information items and having arcs connecting said second information items to said first information items;

transforming said directed graph to a tree-structure graph representing a whole set of said plurality of information items by a root node, subsets of said plurality of information items by branch nodes and respective ones of said plurality of information items by terminal nodes;

segmenting a display region iteratively into a plurality of display positions for each branch node arranged between said root node and said terminal nodes of said tree-structure graph, wherein each segmenting iteration is executed from said root node to said terminal nodes by segmenting a region for a pre-segmented node into several regions which correspond to the branch or terminal nodes connected to said presegmented node; and

displaying each of said plurality of information items at a corresponding one of the positions of the

2. The method for displaying a plurality of information items according to claim 1 wherein the displaying includes displaying said plurality of information items at the each position of the segmented regions according to the attributes of all information items included in each of said plurality of information items.

3. A method for displaying a plurality of information items comprising steps of:

extracting, from said plurality of information items, first information items contained in said plurality of information items and second information items having a high similarity with respect to an attribute of said first information items, for each of said plurality of information items;

setting a distance between said first information items and said second information items such that the first and second information items are arranged closer as the similarity between the first and second information items is higher;

setting display positions of said first information items and said second information items in accordance with said distance;

projecting said first information items and said second information items at points in a display plane such that said distance between said first information items and said second information items is maintained; and

segmenting of said display plane by a Volonoi diagram of projected points.

4. An information display system comprising:

a memory for storing a plurality of information items including display data and attributes thereof;

a display unit, and

a CPU connected to said memory and said display unit for executing a program, the CPU including;

means for extracting, from said plurality of information items, first information items contained in said plurality of information items and second information items having a high similarity with respect to an attribute of said first information items, for each of said plurality of information items stored in said memory;

means for preparing a directed graph having nodes corresponding to said plurality of information items and having arcs connecting said second information items to said first information items;

transform means for transforming said directed graph to a tree-structure graph representing a whole set of said plurality of information items by a root node and respective ones of said plurality of information items by terminal nodes;

segmentation means for segmenting the display region of said display unit iteratively for a branch node between said root node and said terminal nodes of said tree-structure graph wherein each segmenting iteration is executed from said root node to said terminal nodes by segmenting a region for a presegmented node into several regions which correspond to the branch or terminal nodes connected to said presegmented node; and

output means for selectively painting each segmented display region by gray levels, colors, and screen patterns respectively corresponding to each of said plurality of information items and outputting painted display region data to said display unit.

5. The information display system according to claim 4 wherein said plurality of information items are merchandise related information.

6. The information display system according to claim 5 wherein said display region corresponds to a sales floor layout chart of the merchandise.

7. A method for analyzing merchandise information comprising steps of storing data of merchandise and prices and other attributes of merchandise in said memory;

inputting said data of the merchandise at a point of sales of the merchandise;

accumulating said prices to obtain accumulated sales data of the merchandise in response to the input of said data of the merchandise;

storing with accumulated sales data in said memory together with said data of the merchandise;

selecting promising merchandise whose accumulated sales data are more than pre-set values, from said merchandise;

analyzing the promising merchandise in a multi-dimensional space whose dimensions represent said attributes using a mathematical cluster-analysis method;

obtaining merchandise information which describes characteristics of the promising merchandise based on a result of the cluster analysis;

storing attribute data of the merchandise together with the data of the merchandise wherein the attribute data of the merchandise are key words;

selecting the attribute data close to a center of clusters from said attribute data based on the result of said cluster analysis.

8. The method for analyzing merchandise information according to claim 7 wherein said cluster analysis step includes steps of:

determining a set of key words of the promising merchandise;

transforming the set of key words to a set of "0"s and "1"s;

and clustering the set of "0"s and "1"s.

9. A method for analyzing merchandise information containing various attributes by a computer comprising the steps of:

analyzing a plurality of merchandise in a multi-dimensional space whose dimensions represent said attributes using a mathematical cluster-analysis method wherein the attributes of the merchandise are represented by key words;

grouping the plurality of merchandise based on a result of the cluster analysis; and

producing attributes which describes characteristics of respective groups of the plurality of merchandise.