Patent Application
20060005121
This invention relates in general to the field of data analysis and online analytical processing. More particularly, this invention relates to the use of online analytical processing attributes using data mining techniques.
Data Mining
Data mining is a general term for a field of computing in which trends, patterns, and relationships are uncovered from accumulated electronic data. Data mining (sometimes termed “knowledge discovery”) allows the use of a data store by examining the data for patterns, e.g., to suggest better ways to produce profit, savings, higher quality products, and greater customer satisfaction. Data mining is used to sift through large amounts of data and the associated many competing and potentially useful dimensions of analysis and associated combinations.
For example, a business may amass a large collection of information about its customers. This information may include purchasing information and any other information available to the business about the customer. The predictions of a model associated with customer data may be used, for example, to control customer attrition, to perform credit-risk management, to detect fraud, or to make decisions on marketing.
Intelligent cross-selling support may be provided. For example, the data mining functionality may be used to suggest items that a user might be interested in by correlating properties about the user, or items the user has ordered, with a database of items that other users have ordered previously. Users may be segmented based on their behavior or profile. Data mining allows the analysis of segment models to discover the characteristics that partition users into population segments. Additionally, missing values in user profile data may be predicted. For example, where a user did not supply data, the value for that data may be predicted.
Data Warehousing and OLAP
Online analytical processing (OLAP) is a key part of many data warehouse and business analysis systems. OLAP services provide for fast analysis of multidimensional information. For this purpose, OLAP services provide for multidimensional access and navigation of data in an intuitive and natural way, providing a global view of data that can be drilled down into particular data of interest. Speed and response time are important attributes of OLAP services that allow users to browse and analyze data online in an efficient manner. Further, OLAP services typically provide analytical tools to rank, aggregate, and calculate lead and lag indicators for the data under analysis.
In this context, an OLAP cube may be modeled according to a user's perception of the data. The cube may have multiple dimensions, each dimension modeled according to attributes of the data. Typically, there is a hierarchy associated with each dimension. For example, a time dimension can consist of years subdivided into months subdivided into weeks subdivided into days, while a geography dimension can consist of countries subdivided into states subdivided into cities. Dimension members can act as indices for identifying a particular cell or range of cells within the cube.
OLAP services are often used to analytically model data that is stored in a relational database such as, for example, an Online Transactional Processing (OLTP) database. Data stored in a relational database may be organized according to multiple tables with each table having data corresponding to a particular data type. A table corresponding to a particular data type may be organized according to columns corresponding to data attributes
The data stored, for example, may represent the business history of an organization. This historical data is used for analysis. In the case of business history data, the analysis can be used to support business decisions at many levels, from strategic planning to performance evaluation of a discrete organizational unit. Data in a data warehouse is organized to support analysis rather than to process real-time transactions as in online transaction processing systems (OLTP).
Online analytical processing (OLAP) technology enables data warehouses to be used effectively for such analysis, providing rapid responses to iterative complex analytical queries. OLAP uses a multidimensional data model and data aggregation techniques to organize and summarize large amounts of data so the data can be evaluated quickly using online analysis and graphical tools. The answer to a query into historical data often leads to subsequent queries as the analyst searches for answers or explores possibilities. OLAP systems provide the speed and flexibility to support the analyst in real time.
Whereas data warehouses and data marts are the data stores for analysis data, online analytical processing (OLAP) is the technology that enables client applications to efficiently access this data. OLAP provides many benefits to analytical users, for example:
In many cases, OLAP uses a dimensional data scheme in order to effect these benefits. For example, multidimensional OLAP cubes are created from the available data. A cube is a specialized database that is optimized to combine, process, and summarize large amounts of data in order to provide answers to questions about that data in the shortest amount of time. This allows users to analyze, compare, and report on data in order to spot business trends, opportunities, and problems. A cube uses pre-aggregated data instead of aggregating the data at the time the user submits a query. Queries are run against these cubes.
The queries of a user yield dimensions of data, which the user can browse in order to view the responsive data. Dimensions may have dimension attributes, which include information about the members of the dimension. For example, a geographical state dimension will allow the dimension to be browsed by state. Some dimension attributes are categorical. Such attributes categorize the members of the dimension into one of several pre-defined states.
Dimensions may include dimension attributes which are continuous. A continuous dimension attribute is one where the value for the attribute may be anywhere within a range of possible values. For example, one such attribute may correspond to the age of a customer. Associated with the age attribute is a range of possible values for the attribute. As another example, cases may correspond to different machine parts. One continuous attribute in the data set may be the weight of the machine part as expressed in milligrams. Browsing a dimension by a continuous dimension attribute may be impossible or useless to the user, because of the infinite possible values for the continuous dimension attribute.
Another type of continuous dimension attribute may be non-numeric, such as city information. While state information may be seen as a categorical dimension attribute, with 50 possible states, city information may not be limited to a set number of cities. Even where the dimension attribute is not continuous (e.g. because there are only a certain number of states) the browsing of a dimension by a dimension attribute with a great number of possible values may not be useful.
One method in which a dimension may be browsed by a dimension attribute which is continuous is through simple discretization. In discretization, the continuous attribute is divided into a number of states by dividing the possible range for the continuous attribute equally into a fixed number of subranges. Each subrange is treated as a distinct state which may be browsed separately.
However, this brute force solution may not provide the most useful division of data into ranges. For example, a continuous attribute may allow for values between Vmin and Vmax. This range may be divided into ten subranges, R1, R2, though R10. If, however, most of the data for a dimension falls into range R4, then the discretization of the continuous attribute may yield little useful information for the browsing user. If all of the data falls into range R4, then no gain would be realized by performing this discretization.
Thus, there is a need for a way to allow dimension attributes to be used for browsing or examination of a dimension or other collection of data, in a way which allows the information contained in the dimension attribute to be more effectively used.
In OLAP data structures or other data modeling contexts, members of a dimension are grouped by allowing one dimension attribute to be discretized. In order to perform this discretization, the distribution of values for the dimension attribute is examined. Values for the dimension attribute for the entire dimension being examined may be considered to determine this distribution, or an approximate distribution may be obtained by using only sample data from the dimension.
Once this distribution is obtained, it is examined in order to divide the range of the dimension attribute into groups or subranges. These groups are then used as “buckets” for the discretization of the dimension attribute. A number of such subranges/buckets are determined. The dimension attribute can then be treated as a categorical attribute, with the value for the categorical attribute for a dimension member with a specific value for the dimension attribute being equal to the state corresponding to the subrange into which that specific value falls. When the dimension is then browsed in an OLAP context, the resulting categorical attribute data may be used to group the members of the dimension being browsed.
Other embodiments are described below.
The foregoing summary, as well as the following detailed description of presently preferred embodiments, is better understood when read in conjunction with the appended drawings. For the purpose of illustrating the invention, there is shown in the drawings exemplary constructions of the invention; however, the invention is not limited to the specific methods and instrumentalities disclosed. In the drawings:
Overview
In accordance with the invention, the data stored in a dimension for a dimension attribute is examined. The distribution of the values for that dimension attribute is determined. Based on the distribution, the range of values for the dimension attribute is divided into subranges.
The dimension attribute can then be treated as a categorical attribute, with the value for the categorical attribute for a member with a specific value for the dimension attribute being equal to the state corresponding to the subrange into which that specific value falls. When data is used, e.g. for browsing in an OLAP context, the resulting categorical attribute data may be used to group the members of the dimension.
Exemplary Computing Environment
One of ordinary skill in the art can appreciate that a computer or other client or server device can be deployed as part of a computer network, or in a distributed computing environment. In this regard, the present invention pertains to any computer system having any number of memory or storage units, and any number of applications and processes occurring across any number of storage units or volumes, which may be used in connection with the present invention. The present invention may apply to an environment with server computers and client computers deployed in a network environment or distributed computing environment, having remote or local storage. The present invention may also be applied to standalone computing devices, having programming language functionality, interpretation and execution capabilities for generating, receiving and transmitting information in connection with remote or local services.
The invention is operational with numerous other general purpose or special purpose computing system environments or configurations. Examples of well known computing systems, environments, and/or configurations that may be suitable for use with the invention include, but are not limited to, personal computers, server computers, hand-held or laptop devices, multiprocessor systems, microprocessor-based systems, set top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments that include any of the above systems or devices, and the like.
The invention may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The invention may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network or other data transmission medium. In a distributed computing environment, program modules and other data may be located in both local and remote computer storage media including memory storage devices. Distributed computing facilitates sharing of computer resources and services by direct exchange between computing devices and systems. These resources and services include the exchange of information, cache storage, and disk storage for files. Distributed computing takes advantage of network connectivity, allowing clients to leverage their collective power to benefit the entire enterprise. In this regard, a variety of devices may have applications, objects or resources that may utilize the techniques of the present invention.
With reference to
Computer 110 typically includes a variety of computer readable media. Computer readable media can be any available media that can be accessed by computer 110 and includes both volatile and nonvolatile media, removable and non-removable media. By way of example, and not limitation, computer readable media may comprise computer storage media and communication media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CDROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and that can accessed by computer 110. Communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media. The term “modulated data signal” means a signal that has one or more of its characteristics set or changed in such a manner as to encode information in the signal. By way of example, and not limitation, communication media includes wired media such as a wired network or direct-wired connection, and wireless media such as acoustic, RF, infrared and other wireless media. Combinations of any of the above should also be included within the scope of computer readable media.
The system memory 130 includes computer storage media in the form of volatile and/or nonvolatile memory such as read only memory (ROM) 131 and random access memory (RAM) 132. A basic input/output system 133 (BIOS), containing the basic routines that help to transfer information between elements within computer 110, such as during start-up, is typically stored in ROM 131. RAM 132 typically contains data and/or program modules that are immediately accessible to and/or presently being operated on by processing unit 120. By way of example, and not limitation,
The computer 110 may also include other removable/non-removable, volatile/nonvolatile computer storage media. By way of example only,
The drives and their associated computer storage media discussed above and illustrated in
The computer 110 may operate in a networked environment using logical connections to one or more remote computers, such as a remote computer 180. The remote computer 180 may be a personal computer, a server, a router, a network PC, a peer device or other common network node, and typically includes many or all of the elements described above relative to the computer 110, although only a memory storage device 181 has been illustrated in
When used in a LAN networking environment, the computer 110 is connected to the LAN 171 through a network interface or adapter 170. When used in a WAN networking environment, the computer 110 typically includes a modem 172 or other means for establishing communications over the WAN 173, such as the Internet. The modem 172, which may be internal or external, may be connected to the system bus 121 via the user input interface 160, or other appropriate mechanism. In a networked environment, program modules depicted relative to the computer 110, or portions thereof, may be stored in the remote memory storage device. By way of example, and not limitation,
While some exemplary embodiments herein are described in connection with software residing on a computing device, one or more portions of the invention may also be implemented via an operating system, application programming interface (API) or a “middle man” object, a control object, hardware, firmware, etc., such that the methods may be included in, supported in or accessed via all of NET's languages and services, and in other distributed computing frameworks as well.
As described above, attributes may be of different types, such as continuous or categorical. Where a continuous-type dimension attribute is included in a dimension it may be less useful for browsing or other purposes than a categorical-type dimension attribute would be. Additionally, when a categorical-type dimension has too many states, it may be difficult to browse the dimension. Thus, according to the present invention, the values in a dimension for a dimension attribute are examined.
In order to discretize a dimension attribute, the distribution of the values for the dimension attribute is determined. This distribution may either be obtained by determining all values in the dimension for the dimension attribute, or by randomly or pseudo-randomly selecting a sample of values for the dimension attribute from the dimension. The distribution of values is then used in order to select subranges of the range of values for the dimension attribute to use as buckets or states for grouping the members of the dimension.
Where the cases are contained in an OLAP structure, the categorical-type attribute information can be used for allowing browsing through a dimension. Browsing commands are accepted from a user, and these browsing commands allow the use of the categorical-type attribute information to select cases for display to the user.
Producing the Subranges
The number of buckets or states to use for the new categorical attribute may be selected by the user. Alternatively, it may be selected dynamically and automatically based on the distribution and the method used to produce the subranges. Some methods of producing subranges may include a method of determining how many subranges there should be. For example, as described below, the agglomeration clustering method can include a way of reducing the number of clusters until a preferred clustering is reached.
As can be seen in
The division of a distribution of values into groupings with significance (rather than a random division of values) is a mathematical problem which has been and continues to be solved in different ways. Both the selection of the number of buckets or states and the actual division of the distribution into subranges may be done by any process.
K-Means
One method which may be used is by performing single-dimensional clustering using the K-Means algorithm. In one embodiment, in order to perform the K-means algorithm, K random locations L1 to LK are selected along the distribution. The datapoints (the values in the data set or sample in the distribution) are then each assigned to the closest of the K random locations. For any location Ln of the K random locations, then, a group of data points has been assigned to that location Ln. A new location is then determined based on data points assigned to the location. For example, the mathematical average of the data points may be determined, and the new location Ln′ set to that average. A number of iterations are performed. The iterations may end at a predetermined iteration, or when the sum of the movements of the locations for the latest iteration is under a specific threshold.
After the K-means algorithm is performed, the subranges for the categorical attribute may then be determined. The K locations at the end of the iteration are the center of clusters. Therefore, if the final locations arranged from one a beginning of the possible range of values to the end are FL1 through FLK, then the first range would be from one end of the range to (FL1+FL2)/2. The second range would be from (FL1+FL2)/2 through (FL2+FL3)/2. The last range would be from (FLK−1+FLK)/2 to the end of the range. This would produce K subranges for the categorical attribute.
Equal Areas
Another possible method for dividing the distribution into subranges is the equal areas method. In this method, the distribution of values across the population (the data set or sample) is analyzed. Bucket ranges are then created such that the total population is distributed equally across the buckets. Thus, under this method, the area under the curve 300 in
Equal areas may be used to divide a distribution into groups even where the dimension attribute is not a numeric attribute. Thus, where the dimension attribute is geographic, into cities, the dimension attribute may be divided into groups such that each group contains approximately the same number of members.
Thresholds
A third possible method for dividing the distribution into subranges according to one embodiment of the invention is by identifying inflection points in the distribution curve. These points correspond to gradient changes from positive to negative. For example, the two subrange boundary lines 400 in
Agglomeration Clustering
A fourth possible method for dividing the distribution into subranges according to one embodiment of the invention is agglomeration clustering. Each case is initially assigned its own cluster. Then, iteratively, all pairs of groups are evaluated and the pair that is closest is found. Closeness may be determined, for example, by comparing the average value in the clusters. The closest clusters are merged into a single group.
This process continues until either the closest clusters do not meet some guideline. For example, if the closest clusters are not close enough, using a threshold value, they may not be merged. Alternatively, the process may continue until a predetermined required number of clusters are achieved. Subranges are then selected so that the values in each cluster are all assigned to one subrange.
Data Set Conversion
Naming of Groups/Buckets
In order to enhance user convenience when dealing with the groups which have been created, the groups or buckets which have been generated may be named. Group or bucket name generation may be performed according to the following grammar:
A template string could be applied which allows the buckets to be named using this grammar. For example, where the template string is: .
.
There are multiple ways of implementing the present invention, e.g., an appropriate API, tool kit, driver code, operating system, control, standalone or downloadable software object, etc. which enables applications and services to use the product configuration methods of the invention. The invention contemplates the use of the invention from the standpoint of an API (or other software object), as well as from a software or hardware object that communicates in connection with product configuration data. Thus, various implementations of the invention described herein may have aspects that are wholly in hardware, partly in hardware and partly in software, as well as in software.
As mentioned above, while exemplary embodiments of the present invention have been described in connection with various computing devices and network architectures, the underlying concepts may be applied to any computing device or system in which it is desirable to implement product configuration. Thus, the techniques for encoding/decoding data in accordance with the present invention may be applied to a variety of applications and devices. For instance, the algorithm(s) and hardware implementations of the invention may be applied to the operating system of a computing device, provided as a separate object on the device, as part of another object, as a reusable control, as a downloadable object from a server, as a “middle man” between a device or object and the network, as a distributed object, as hardware, in memory, a combination of any of the foregoing, etc. While exemplary programming languages, names and examples are chosen herein as representative of various choices, these languages, names and examples are not intended to be limiting. With respect to embodiments referring to the use of a control for achieving the invention, the invention is not limited to the provision of a .NET control, but rather should be thought of in the broader context of any piece of software (and/ore hardware) that achieves the configuration objectives in accordance with the invention. One of ordinary skill in the art will appreciate that there are numerous ways of providing object code and nomenclature that achieves the same, similar or equivalent functionality achieved by the various embodiments of the invention. The term “product” as utilized herein refers to products and/or services, and/or anything else that can be offered for sale via an Internet catalog. The invention may be implemented in connection with an on-line auction or bidding site as well.
As mentioned, the various techniques described herein may be implemented in connection with hardware or software or, where appropriate, with a combination of both. Thus, the methods and apparatus of the present invention, or certain aspects or portions thereof, may take the form of program code (i.e., instructions) embodied in tangible media, such as floppy diskettes, CD-ROMs, hard drives, or any other machine-readable storage medium, wherein, when the program code is loaded into and executed by a machine, such as a computer, the machine becomes an apparatus for practicing the invention. In the case of program code execution on programmable computers, the computing device will generally include a processor, a storage medium readable by the processor (including volatile and non-volatile memory and/or storage elements), at least one input device, and at least one output device. One or more programs that may utilize the product configuration techniques of the present invention, e.g., through the use of a data processing API, reusable controls, or the like, are preferably implemented in a high level procedural or object oriented programming language to communicate with a computer system. However, the program(s) can be implemented in assembly or machine language, if desired. In any case, the language may be a compiled or interpreted language, and combined with hardware implementations.
The methods and apparatus of the present invention may also be practiced via communications embodied in the form of program code that is transmitted over some transmission medium, such as over electrical wiring or cabling, through fiber optics, or via any other form of transmission, wherein, when the program code is received and loaded into and executed by a machine, such as an EPROM, a gate array, a programmable logic device (PLD), a client computer, a video recorder or the like, or a receiving machine having the signal processing capabilities as described in exemplary embodiments above becomes an apparatus for practicing the invention. When implemented on a general-purpose processor, the program code combines with the processor to provide a unique apparatus that operates to invoke the functionality of the present invention. Additionally, any storage techniques used in connection with the present invention may invariably be a combination of hardware and software.
While the present invention has been described in connection with the preferred embodiments of the various figures, it is to be understood that other similar embodiments may be used or modifications and additions may be made to the described embodiment for performing the same function of the present invention without deviating therefrom. For example, while exemplary network environments of the invention are described in the context of a networked environment, such as a peer to peer networked environment, one skilled in the art will recognize that the present invention is not limited thereto, and that the methods, as described in the present application may apply to any computing device or environment, such as a gaming console, handheld computer, portable computer, etc., whether wired or wireless, and may be applied to any number of such computing devices connected via a communications network, and interacting across the network. Furthermore, it should be emphasized that a variety of computer platforms, including handheld device operating systems and other application specific operating systems are contemplated, especially as the number of wireless networked devices continues to proliferate. Still further, the present invention may be implemented in or across a plurality of processing chips or devices, and storage may similarly be effected across a plurality of devices. Therefore, the present invention should not be limited to any single embodiment, but rather should be construed in breadth and scope in accordance with the appended claims.
