Claims:1. A method for extracting relevant data fields from a set of documents, the method comprising a processor implemented steps of:

collating a training data obtained by an optical character recognition tool and manually extracted data (202);
identifying location of a plurality of spatial cues corresponding to a plurality of data fields in a document, wherein the document is part of the training data (206);
creating a list of keywords used to locate a specific data field for the document from the training data (208);
creating a list of data fields that occurred proximate to each other in a predefined sequence specific for the document from the training data (210);
marking content of the document as either relevant or non-relevant based on the occurrences of data field values in the document, wherein the location of the plurality of spatial cues, the list of keywords, the list of data fields and the content markings referred as domain specific knowledge (212);
computing conditional random fields to generate a training model using domain specific knowledge as inputs (214);
providing the set of documents as an input to the training model (216); and
extracting the relevant data fields from the set of documents using the training model (218).

2. The method of claim 1 further comprising the step of applying under-sampling on the training data.

3. The method of claim 1, wherein the conditional random fields are a class of discriminative models that predicts a global probability of a sequence of random variables assuming first order Markov dependency between a set of variables.

4. The method of claim 1 wherein the training model is shared with a user in the form of an extractor utility.

5. The method of claim 1 further comprising the step of providing a mapping logic to spot an excel data from text files in the set of documents.

6. The method of claim 1, wherein the list of keywords is constructed from up to 5 previous observations and a keyword or phrase is located within the list.

7. The method of claim 1, wherein the step of marking the content as relevant or non-relevant further comprises:

splitting text of the document into sentences using a heuristics-based sentence model;
constructing a logistic regression-based classifier with n-gram features by locating the data fields that were manually extracted in sentences of the text; and
marking the training data for this classifier sentences as relevant or non-relevant based on the occurrences of the data field values.

8. The method of claim 1, wherein the set of documents are collated from the mortgage domain.

9. The method of claim 1 further comprising the step of cleaning and syncing the collated training data (204).

10. A system for extracting relevant data fields from a set of documents, the system comprising:

a training data storage module (102) storing training data, wherein the training data collated with data obtained from an optical character recognition tool (124) and the manually extracted data;
a memory (106);
and a processor (108) in communication with the memory, the processor further comprising:
a synching module (110) for cleaning and syncing the training data;
a spatial context identification module (112) for identifying location of a plurality of spatial cues corresponding to a plurality of data fields in a document, wherein the document is part of the training data;
a textual similarity context identification module (114) for creating a list of keywords used to locate a specific data field for the document from the training data;
a chain frequency context identification module (116) for creating a list of data fields that occurred proximate to each other in a predefined sequence specific for the document from the training data;
a relevancy module (118) for marking content of the document as either relevant or non-relevant based on the occurrences of data field values in the document, wherein the location of the plurality of spatial cues, the list of keywords, the list of data fields and the content markings referred as domain specific knowledge;
a CRF computation module (120) for computing a conditional random fields to generate a training model using the domain specific knowledge as inputs;
an input module (104) for providing the set of documents as an input to the training model; and
an extraction module (122) for extracting the relevant data fields from the set of documents using the training model.
, Description:FORM 2

THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENT RULES, 2003

COMPLETE SPECIFICATION
(See Section 10 and Rule 13)

Title of invention:
METHOD AND SYSTEM FOR EXTRACTING DATA FIELDS FROM DOCUMENTS USING CONDITIONAL RANDOM FIELDS

Applicant:
Tata Consultancy Services Limited
A company Incorporated in India under the Companies Act, 1956
Having address:
Nirmal Building, 9th Floor,
Nariman Point, Mumbai 400021,
Maharashtra, India

The following specification particularly describes the invention and the manner in which it is to be performed.

TECHNICAL FIELD
[001] The embodiments herein generally relates to the field of information extraction from a set of documents, and, more particularly, to a method and system for extracting relevant data fields from the set of documents.

BACKGROUND
[002] With a load of documents being generated every day. The relevant information extraction from these documents has become an important field of research. Information extraction from different types of documents has always been a challenging task. The scale of this task was made even more challenging by the variations of different kinds that occur. A system is needed that would efficiently address the variations, in contrast to template-based data field identification, extraction, and entry, which is the prevalent practice for this task.
[003] Few industries for example, the mortgage industry presents unique challenges with regards extracting data from key documents. A typical loan file may consist of hundreds of pages of documents. Manually referring to these documents, often as scanned images, to extract data fields is a laborious, time-consuming task.
[004] Manually extracting data from scanned documents requires considerable expertise, is time consuming, and yet can be quite error prone. To (semi-) automate the extraction process, it is possible to process the text of the documents using as source either the portable document format (PDF) in which scanned documents are stored, or the text obtained from PDF documents using optical character recognition (OCR). Existing technologies used to extract information from PDF/text documents face problems in terms of being unable to address the multitude of variations in text formatting, especially when extracting information from large number of documents, extracting 30-60 fields from a document for close to 100 document types, such that the size of overall text of the documents is much larger than the size of text where the mentions of fields actually occur.

SUMMARY
[005] The following presents a simplified summary of some embodiments of the disclosure in order to provide a basic understanding of the embodiments. This summary is not an extensive overview of the embodiments. It is not intended to identify key/critical elements of the embodiments or to delineate the scope of the embodiments. Its sole purpose is to present some embodiments in a simplified form as a prelude to the more detailed description that is presented below.
[006] In view of the foregoing, an embodiment herein provides a system for extracting relevant data fields from a set of documents, the system comprises a training data storage module, a memory and a processor. The training data storage module storing training data, wherein the training data collated with data obtained from an optical character recognition tool and the manually extracted data. The processor further comprises a synching module, a spatial context identification module, a textual similarity context identification module, a chain frequency context identification module, a relevancy module, a CRF computation module, an input module and an extraction module. The synching module cleans and syncs the training data. The spatial context identification module identifies location of a plurality of spatial cues corresponding to a plurality of data fields on a page of a document, wherein the document is part of the training data. The textual similarity context identification module creates a list of keywords used to locate a specific data field for the document from the training data. The chain frequency context identification module creates a list of data fields that occurred proximate to each other in a predefined sequence specific for the document from the training data; The relevancy module marks content of the document as either relevant or non-relevant based on the occurrences of data field values in the document, wherein the location of the plurality of spatial cues, the list of keywords, the list of data fields and the content markings referred as domain specific knowledge. The CRF computation module computes a conditional random fields to generate a training model using the domain specific knowledge as inputs. The input module provides the set of documents as an input to the training model. The extraction module extracts the relevant data fields from the set of documents using the training model.
[007] In another aspect the embodiment here provides a method for extracting relevant data fields from a set of documents. Initially, a training data obtained by an optical character recognition tool and manually extracted data is collated. In the next step, the location of a plurality of spatial cues corresponding to a plurality of data fields are identified on a page of a document, wherein the document is part of the training data. Further, a list of keywords used to locate a specific data field for the document from the training data is created. In the next step, a list of data fields is created that occurred in the vicinity of each other in a predefined sequence specific for the document from the training data; In the next step, content of the document is marked as either relevant or non-relevant based on the occurrences of data field values in the document, wherein the location of the plurality of spatial cues, the list of keywords, the list of data fields and the content markings referred as domain specific knowledge. In the next step, a conditional random fields is computed to generate a training model using domain specific knowledge as inputs. The set of documents are then provided as an input to the training model. And finally, the relevant data fields are extracted from the set of documents using the training model.
[008] It should be appreciated by those skilled in the art that any block diagram herein represent conceptual views of illustrative systems embodying the principles of the present subject matter. Similarly, it will be appreciated that any flow charts, flow diagrams, state transition diagrams, pseudo code, and the like represent various processes which may be substantially represented in computer readable medium and so executed by a computing device or processor, whether or not such computing device or processor is explicitly shown.

BRIEF DESCRIPTION OF THE DRAWINGS
[009] The embodiments herein will be better understood from the following detailed description with reference to the drawings, in which:
[010] Fig. 1 illustrates a block diagram of a system for extracting relevant data fields from a set of documents according to an embodiment of the present disclosure;
[011] Fig. 2 shows a schematic representation of the system for extracting relevant data fields from a set of documents according to an embodiment of the disclosure; and
[012] Fig. 3A-3B is a flowchart illustrating the steps involved in extracting relevant data fields from a set of documents according to an embodiment of the present disclosure.

DETAILED DESCRIPTION
[013] The embodiments herein and the various features and advantageous details thereof are explained more fully with reference to the non-limiting embodiments that are illustrated in the accompanying drawings and detailed in the following description. The examples used herein are intended merely to facilitate an understanding of ways in which the embodiments herein may be practiced and to further enable those of skill in the art to practice the embodiments herein. Accordingly, the examples should not be construed as limiting the scope of the embodiments herein.
[014] Referring now to the drawings, and more particularly to Fig. 1 through Fig. 3, where similar reference characters denote corresponding features consistently throughout the figures, there are shown preferred embodiments and these embodiments are described in the context of the following exemplary system and/or method.
[015] According to an embodiment of the disclosure, a system 100 for extracting relevant data fields from a set of documents is shown in Fig. 1. A schematic representation of the system 100 is shown in Fig. 2. The system 100 uses conditional random fields (CRFs), a discriminative model that predicts global probability of sequence of random variables, wherein the probability structure depends on a log-linear combination of evidence features derived from the input text. The system 100 is a machine learning, natural language processing system. The system 100 can significantly be used in the mortgage industry as explained with the help of example below. Though the use of system 100 to extract data from any other industry / domain is well within the scope of this disclosure. Therefore the disclosure have been explained generically and later explained with the help of the example.
[016] According to an embodiment of the disclosure, the system 100 further comprises a training data storage module 102, an input module 104, a memory 106 and a processor 108 as shown in the block diagram of Fig. 1. The processor 108 works in communication with the memory 106. The processor 108 further comprises a plurality of modules. The plurality of modules accesses the set of algorithms stored in the memory 106 to perform a specific task. The processor 108 further comprises a synching module 110, spatial context identification module 112, a textual similarity context identification module 114, a chain frequency context identification module 116, a relevancy module 118, a conditional random fields (CRF) computation module 120 and an extraction module 122.
[017] According to an embodiment of the disclosure, the training data is stored in the training data storage module 102. The training data collated with data obtained from an optical character recognition tool 124 and the manually extracted data. The optical character recognition tool scan various documents to generate the training data. The manually extracted data is generally stored in an Excel file. The manually extracted data is extracted manually by the users working on the documents.
[018] The training data is provided to the synching module 110. The synching module 110 is configured to perform cleaning and syncing between text documents (identified by numbers) and the Excel data (in which a data field called docNbr represent the document numbers) is required before training. Additionally, the system 100 also comprises a mapping module 126. The mapping module 126 facilitates a mapping logic required (coded once) to spot the Excel data in the text documents.
[019] According to an embodiment of the disclosure the input module 104 is configured to provide an input to the system 100. The input module 104 is configured to provide the set of documents as input to the training model as explained in the later part of the disclosure. The input module 104 can include a variety of software and hardware interfaces, for example, a web interface, a graphical user interface, and the like and can facilitate multiple communications within a wide variety of networks N/W and protocol types, including wired networks, for example, LAN, cable, etc., and wireless networks, such as WLAN, cellular, or satellite.
[020] According to an embodiment of the disclosure, the system 100 uses various domain independent aspect and domain specific aspects. The system 100 learns contextual features, which mimic how a human uses approximate contexts, i.e., text spans, either before and after or around the mentions of specific entity types, to recognize them. To ensure that system learns to generalize across varying contexts, the contexts are learnt rather than specified beforehand.
[021] To map the extracted fields to their occurrences in the text, we use domain-specific patterns were used. Certain entities specific to the domain occur in certain ways in the text. We use clustering to find out highly occurring keywords, which indicate the presence of a mention of a specific entity.
[022] The system 100 comprises the spatial context identification module 112 as shown in Fig. 1. The spatial context identification module 112 is configured to identify the location of a plurality of spatial cues corresponding to a plurality of data fields on a page of a document, wherein the document is part of the training data. The plurality of spatial cues are spatial arrangement of the data fields in the scanned document generated from the optical character recognition tool 124. Such as a specific data field tends to occur, whether at the top right of the first page, or at the bottom of the second page, or some other specific place in the document. Thus spatial cues were taken into consideration. In an example, this is done by imagining slots, in which the start index of a particular data field would fall. While training, the text is divided so as to make slots of 25 characters. Depending on the document type, the data fields at the top right of the scanned PDF would be (Optical character recognition) OCRed as 1st slot in the text. This simple representation enables capturing the spatial arrangement of all data fields of a document type in terms of numbered slots. This feature is implemented by first computing the top 4 most frequent slots for each data field for a given dataset containing number of documents and then during training, checking whether the slot of the previous observation falls within the top slots found for the current label.
[023] According to an embodiment of the disclosure, the system 100 comprises the textual similarity context identification module 114 as shown in Fig. 1. The textual similarity context identification module 114 is configured to create a list of keywords used to locate a specific data field for the document from the training data. The list of keywords is maintained which is used to locate the specific data field for a given document type. This feature is implemented by maintaining a list of data fields and the corresponding keywords. In an example, a string is constructed from up to 5 previous observations and a keyword or phrase is located within the string. The feature is set when the current label corresponds with the data field associated with the keyword or the phrase. The list of keywords currently used includes not only the keywords that the associates use but also other keywords and phrases that were found by clustering the occurrences of the data fields including 30 characters to the left and the right of each occurrence. This was mainly the case for the data fields for which the associates did not maintain a keyword list. An example of the textual keywords used to locate specific data fields is shown in Table 1 below.

[024] According to an embodiment of the disclosure, the system 100 further comprises the chain frequency context identification module 116 as shown in Fig. 1. The chain frequency context identification module 116 creates a list of data fields that occurred in the vicinity of each other in a predefined sequence specific for the document from the training data. The chain frequency were specific to a kind of document, i.e., a document from a certain state, a certain county, and some attorneys. This feature is implemented by computing the most frequent set of next data fields for a given data field in the training data. The feature is set when the set of next frequent labels for the previous label matches with the current label.
[025] According to an embodiment of the disclosure, the system 100 further comprises the relevancy module 118 as shown in Fig. 1. The relevancy module 118 marks content of the document as either relevant or non-relevant based on the occurrences of data field values in the document, To classify the content of the document into relevant and non-relevant, the text of the document is first split into sentences using a heuristics-based sentence model. A logistic regression-based classifier is constructed with n-gram features by locating the data fields that were manually extracted in the sentences of the text. The training data for this classifier is therefore sentences (i.e., blocks of text of configurable length) marked as relevant or non-relevant based on the occurrences of the data field values. Since the text of the documents may or may not be amenable for sentence detection, an initial heuristics-based sentence detection was carried out, and then split the text in blocks of configurable length.
[026] Thus, the location of the plurality of spatial cues, the list of keywords, the list of data fields and the content markings are referred as the domain specific knowledge. The domain specific knowledge is now used as input to generate a training model as shown in Fig. 2.
[027] According to an embodiment of the disclosure, the system 100 further comprises the conditional random fields (CRF) computation module 120, the input module 104 and the extraction module 122. The CRF computation module 120 computes a conditional random fields to generate the training model using the domain specific knowledge as inputs. Once the training is over, the trained model is shared as an extractor utility to the user. The user provide the set of documents as an input to the training model using the input module 104. The extraction module 122 which comprises the training model extracts the relevant data fields from the set of input documents.
[028] The conditional random fields (CRFs) computation is a discriminative model that predicts global probability of sequence of random variables, wherein the probability structure depends on a log-linear combination of evidence features derived from the input.
[029] In operation, a flowchart 200 illustrating the steps of extracting relevant data fields from the set of documents is shown in Fig. 3A-3B. Initially at step 202, the training data is collated. The training data is obtained by the optical character recognition tool 124 and manually extracted data. The optical recognition tool 124 obtain data from various format of such as PDF file etc. While the manually extracted data is obtained from the Excel file. At step 204, the training data is cleaned and synched.
[030] In the next step 206, the location of the plurality of spatial cues corresponding to a plurality of data fields on a page of the document is identified, wherein the document is part of the training data collated in the first step. At step 208, the list of keywords used to locate the specific data field for the document from the training data is created. At step 210, the list of data fields that occurred in the vicinity of each other in a predefined sequence specific for the document from the training data is also created. It should be appreciated that the step 206 to 210 were decided to perform based on some observation of the team who was manually working on the training data. Further at step 212, the content of the document is marked as either relevant or non-relevant based on the occurrences of data field values in the document. So the location of the plurality of spatial cues identified in step 206, the list of keywords created at step 208, the list of data fields created at step 210 and the content markings marked at step 212 are referred as domain specific knowledge.
[031] In the next step 214, the conditional random fields (CRF) are created to generate the training model using the domain specific knowledge as inputs. At step 216, the set of documents are provided as the input to the training model. And finally at step 218, the relevant data fields are extracted from the set of documents using the training model generated in step 214.
[032] According to an embodiment of the disclosure, the system 100 for extracting relevant data fields from the set of documents can also be explained with the help of following example. The set of documents were taken from the field of mortgage domain. Several experiments were conducted with the datasets consisting of subsets of 500 documents each of the 3 major document types, namely Release of Mortgage, Deed of trust, and Deed types. The domain specific knowledge were arrived at gradually as the research team observed the associates and implemented the knowledge.
[033] Table 2 shows the results of the application of the CRF system to all of 500 documents each for the 3 major document types, with 10-fold cross validation.

TABLE 2: Comparison of the performance of different features combinations for multi- and single county mortgage documents
[034] Domain-specific Features: The base system is the one that uses CRF features that have been shown to be effective in sequence labeling, specifically NER, namely the current, previous and next tokens (observations), Part of Speech tag of the current, previous, and next tokens, shape of the current, previous, and next tokens3, and up to 4 tokens to the left and right of the current token. Following the base system, are the systems with the features described earlier, namely frequency cues, spatial cues, and textual context cues, individually, and finally all combined.
[035] The results of the experiments are shown in TABLE 2, per 50 epochs, for the documents from different counties in the document set for all 3 document types and also documents from a single county in the document set for 2 out of 3 document types, albeit only for the systems with combined features. In the experiments, the field value must match completely, i.e., the entire span of the value, as in the CoNLL-2003 shared task.
[036] Under-sampling and Two-stage Classification: Prior to implementing the relevancy classifier, the recall of the system suffered considerably. Upon implementing the relevancy classifier, it was found that in case of the Deed of trust documents, the relevant content was only about 5-10% of the entire text. The same held true for the Deed documents. Considering that majority of 84 and 52 fields are to be extracted from this small relevant content, a CRF model trained on all of the content would perform poorly due to all the noise (non-relevant content). To differentiate the performance of the CRF system without and with relevancy under-sampling and balanced under-sampling for training data, experiments with another set of 500 documents were performed for the document types Deed and Release of Mortgage. Additionally, two-stage classification was also performed for this dataset by the procedure described in the previous section. These results are shown in Table 3.
TABLE 3: Under-sampling and two stage classification
Evaluation of experiment:
[037] The overall approach seems to perform considerably better than the template-based approach, which achieves an F1 scores in the range of 10%-15%. The CRF system, here all variants of it, performs better for the Release of Mortgage documents, which constitute the largest document type by volume with an F1 score more than 70.
[038] Feature-specific evaluation: The system with combined features performs better for the document types Deed of trust and Deed, in a single county scenario, rather than when the document set contains documents from multiple counties. This is attributed to the fact that the documents from multiple counties introduce county-specific variations, as described earlier in Section 2, as opposed to the documents from a single county. Specific to the features, following observations were made:
[039] Spatial Cues/Slots- The spatial context features increase the precision for the Release of Mortgage documents as opposed to the Deed of trust and Deed documents, for which the recall increases. This is attributed to spatial consistency in the Release of Mortgage documents, and spatial variations in the Deed of trust and Deed documents. For the Deed of trust and Deed documents, the data fields do not follow a specific spatial arrangement for key fields, especially across the documents from multiple counties.
[040] Textual Context Cues- These features tend to increase recall for all the 3 document types. In most of the experiments, it was found that keywords indicating specific data fields are consistent across all 3 document types, irrespective of whether the documents are from multiple counties or a single county.
[041] Chain Frequency Cues- These features, like textual context features tend to increase recall for all the 3 document types, due to long range dependencies captured in the chain frequencies from the training data.
[042] Under-sampling and Two-stage Classification Table 3 shows the results for applying relevancy, balanced under-sampling, and two-stage classification to a CRF system with all the domain-specific features. It can be seen that relevancy under-sampling is a critical component, especially for Mortgage data extraction. Without the relevancy under-sampling, applied to both training and testing data, the CRF system is restricted in its applicability by very low recall values. The F1 scores for features plus relevancy are comparable to the scores of the same system against different dataset as shown in earlier in Table 2.
[043] With a balanced under-sampling ratio of 2:3, the results show considerable reduction in precision with comparable recall values. The most obvious cause of decrease in precision is that discriminating context is lost by the removal of tokens with O tag from the training data, but not from the testing data. Whereas relevancy under-sampling increases the recall due to its applicability to both the training and the testing data, the balanced under-sampling reduces precision. It was determined that in the context of Mortgage data extraction, balanced under-sampling or any other form of under-sampling applicable only on the training data is not suitable. It can be seen that two-stage classification achieves minor improvements in precision values while more or less maintaining the recall values. Given the unique volumes in the Mortgage domain, the idea of using two-stage classifiers was put on hold, although we continue to explore other ways of including long range dependencies into the CRF system.
[044] Concluding remarks on experiments Apart from being imbalanced with respect to the NONE class, Table 4 shows that Mortgage data is also imbalanced from the point of view of various data fields. Although only 11 top fields in a set of 500 files for each document set is shown, for Deed and Deed of trust document types, many fields appear in less than 100 documents. The CRF system with the domain-specific features and relevancy under-sampling is determined to be the best solution within the various combinations presented.

TABLE 4: Document-wise data field occurrence in a set of 500 files each

[045] Based on these observations, it was decided that the extractor utility for the on ground associates would be formed. The extractor utility extracts fields and store them in a file from which it is uploaded to the workstations of the users. The extractor utility is run on held-out data to find out if the trained model results in 0 true positives for some fields. A list of such fields found is maintained, so that when the extractor utility is run on the incoming documents, predictions pertaining to these fields are hidden and shown only if the maker requests them.
[046] The written description describes the subject matter herein to enable any person skilled in the art to make and use the embodiments. The scope of the subject matter embodiments is defined by the claims and may include other modifications that occur to those skilled in the art. Such other modifications are intended to be within the scope of the claims if they have similar elements that do not differ from the literal language of the claims or if they include equivalent elements with insubstantial differences from the literal language of the claims.
[047] The embodiments of present disclosure herein provides a method and system for extracting relevant data fields from a set of documents. The method thus treat the problem of relevant information extraction from multiple types of documents with precision.
[048] It is, however to be understood that the scope of the protection is extended to such a program and in addition to a computer-readable means having a message therein; such computer-readable storage means contain program-code means for implementation of one or more steps of the method, when the program runs on a server or mobile device or any suitable programmable device. The hardware device can be any kind of device which can be programmed including e.g. any kind of computer like a server or a personal computer, or the like, or any combination thereof. The device may also include means which could be e.g. hardware means like e.g. an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a combination of hardware and software means, e.g. an ASIC and an FPGA, or at least one microprocessor and at least one memory with software modules located therein. Thus, the means can include both hardware means and software means. The method embodiments described herein could be implemented in hardware and software. The device may also include software means. Alternatively, the embodiments may be implemented on different hardware devices, e.g. using a plurality of CPUs.
[049] The embodiments herein can comprise hardware and software elements. The embodiments that are implemented in software include but are not limited to, firmware, resident software, microcode, etc. The functions performed by various modules described herein may be implemented in other modules or combinations of other modules. For the purposes of this description, a computer-usable or computer readable medium can be any apparatus that can comprise, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device.
[050] The medium can be an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system (or apparatus or device) or a propagation medium. Examples of a computer-readable medium include a semiconductor or solid state memory, magnetic tape, a removable computer diskette, a random access memory (RAM), a read-only memory (ROM), a rigid magnetic disk and an optical disk. Current examples of optical disks include compact disk-read only memory (CD-ROM), compact disk-read/write (CD-R/W) and DVD.
[051] A data processing system suitable for storing and/or executing program code will include at least one processor coupled directly or indirectly to memory elements through a system bus. The memory elements can include local memory employed during actual execution of the program code, bulk storage, and cache memories which provide temporary storage of at least some program code in order to reduce the number of times code must be retrieved from bulk storage during execution.
[052] Input/output (I/O) devices (including but not limited to keyboards, displays, pointing devices, etc.) can be coupled to the system either directly or through intervening I/O controllers. Network adapters may also be coupled to the system to enable the data processing system to become coupled to other data processing systems or remote printers or storage devices through intervening private or public networks. Modems, cable modem and Ethernet cards are just a few of the currently available types of network adapters.
[053] A representative hardware environment for practicing the embodiments may include a hardware configuration of an information handling/computer system in accordance with the embodiments herein. The system herein comprises at least one processor or central processing unit (CPU). The CPUs are interconnected via system bus to various devices such as a random access memory (RAM), read-only memory (ROM), and an input/output (I/O) adapter. The I/O adapter can connect to peripheral devices, such as disk units and tape drives, or other program storage devices that are readable by the system. The system can read the inventive instructions on the program storage devices and follow these instructions to execute the methodology of the embodiments herein.
[054] The system further includes a user interface adapter that connects a keyboard, mouse, speaker, microphone, and/or other user interface devices such as a touch screen device (not shown) to the bus to gather user input. Additionally, a communication adapter connects the bus to a data processing network, and a display adapter connects the bus to a display device which may be embodied as an output device such as a monitor, printer, or transmitter, for example.
[055] The preceding description has been presented with reference to various embodiments. Persons having ordinary skill in the art and technology to which this application pertains will appreciate that alterations and changes in the described structures and methods of operation can be practiced without meaningfully departing from the principle, spirit and scope.