FORM 2
THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENT RULES, 2003
COMPLETE SPECIFICATION
(See Section 10 and Rule 13)
Title of invention: SYSTEM AND METHOD FOR ENTITY SELECTION USING AN IVR SYSTEM
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.

FIELD OF INVENTION
The present invention relates to an entity selection over an interactive communication network. More particularly, the invention relates to a system and method for selecting entity while accessing catalogue using IVR system.
BACKGROUND OF THE INVENTION
Now days, Interactive Voice Response systems are widely used due to their easily reachable infrastructure and plenty of information with respect to a user's concerns. The use of IVR's has now extended to world of e-commerce which are engaged in transaction of products such as buying and selling through e-commerce channel is increasing. E-commerce catalogue comprises of enormous number of products from which user can select products of their choice either to buy the products or list the products which is for sale through e-commerce channel. Other than IVR's these e-commerce product catalogues can be accessed by using the internet.
If the catalogues are accessed through internet, then selection of product becomes easy as a user has to simply select a product by entering the product reference number mentioned on the shopping portal. While, if the user is accessing the catalogue using IVR, he has an option to enter the product choice by pressing the number. The existing Interactive Voice Response (IVR) system gives a catalogue with a list of items to select from and the user presses the number corresponding to the item. The current IVR system does not have support for online shopping as in web based shopping systems. The IVR system faces challenges when the number of menu items is more than nine. The categories of items available on web based online shopping portals are not be used in an IVR based systems due to size constraints. If a user wants to go beyond nine menu items, the user faces limitations of number of keys present on the keypad. Thus, the users are restricted for the items they can choose from and the IVR system developers cannot create more than nine menu items. In addition to this, current IVR system does not allow users to enter item name if the item is not available in the catalogue.

Other than this, the present IVR systems do not provide a comfort of accessing the catalogue in a user's native language. The present IVR systems also lacks in switching from one language to another as they lacks in covering plurality of languages for selecting a product while shopping remotely.
Although, other methods are also there where character by character mapping is used, but those methods are not suggestive in nature and moreover, a user has to enter the complete text.
Thus the present IVR systems are time consuming and moreover have a restriction in terms of number of digits and language if a user has to select an option by pressing a number.
Therefore, there is a need to provide a system and method which allows the user to select products from the product catalogue which includes more items in the catalogue. Also, the user should be able to access the product catalogue which accepts product names in the system and suggest relevant products to the user in his native language.
OBJECTS OF THE INVENTION
It is the primary object of the present invention to provide a system and method facilitating an entity selection over an interactive communication network.
It is another object of the system to provide an entity selection in numerical form to be mapped with entity names thus stored in the system.
It is another object of the invention to filter the mapped entities for retrieving closest results with respect to a user's numerical selection.
It is another object of the invention to facilitate selection of language in order to select the entity in a user's native language.
It is another object of the invention to enable the user to access the product catalogue which accepts product names in the system and suggest relevant products to the user.

SUMMARY OF THE INVENTION
The present invention provides a system facilitating an entity selection over an interactive communication network. The system comprises a communication device configured to provide an access to a user to one or more catalogue containing a list of entities. Further, the communication device upon successful access further configured to select an entity in a numerical form such that the numerical form is a representation of the entity name. The system further comprises an interactive voice system communicating with the communication device, configured to process the numerical selection in order to retrieve one or more closest results by constructing the entity name with respect to the numerical entity selection. The interactive voice system further comprises a location identification module configured to identify a location of the communication device in order to select the language for the catalogue thus accessed by the user. The interactive voice system further comprises an entity mapping engine configured to correlate the numerical selection with a set of pre-identified numerical selections for a particular entity and map the correlated numerical selections with the one or more corresponding predefined entity names thus stored in a database. The interactive voice response system further comprises a result filtration module configured to apply a selection mechanism to sort the irrelevant results from the mapped entity results and select a predetermined number of closest results with respect to the user's entity selection. The system for entity selection further comprises an output generation module configured to accept the user's entity selection from the closest mapped results and validate the user's selection for one or more utility purpose.
The present invention also provides a method facilitating an entity selection over an interactive communication network. The method comprises steps of providing an access to a user to one or more catalogue containing list of entities to and selecting the entity in a numerical form, such that numerical form is a representation of the entity name. The method further comprises steps of processing the numerical selection in order to retrieve one or more closest results by constructing the entity name with respect to the numerical entity selection. The processing of numerical selection further comprises steps of

identifying a location of the user in communication in order to select the language for the catalogue thus accessed by the user. The processing of numerical selection further comprises steps of correlating the numerical selection with a set of pre-identified numerical selections for a particular entity and map the correlated numerical selections with the one or more corresponding predefined entity names thus stored in a database. The processing of numerical selection further comprises steps of applying a selection mechanism to sort the irrelevant results from the mapped entity results and select a predetermined number of closest results with respect to the user's entity selection. The method facilitating an entity selection over an interactive communication network further comprises accepting the user's entity selection from the closest mapped results and validate the user's the user's selection for one or more utility purpose.
BRIEF DESCRIPTION OF DRAWINGS
Figure 1 illustrates the system architecture for selecting an entity in accordance with an embodiment of the invention.
Figure 2 illustrates the method of entity selection in accordance with an exemplary embodiment of the system.
Figure 3 illustrates the flow chart for entity selection in accordance with an exemplary embodiment of the invention.
DETAILED DESCRIPTION
Some embodiments of this invention, illustrating its features, will now be discussed:
The words "comprising", "having", "containing", and "including", and other forms thereof, are intended to be equivalent in meaning and be open ended in that an item or items following any one of these words is not meant to be an exhaustive listing of such item or items, or meant to be limited to only the listed item or items.

It must also be noted that as used herein and in the appended claims, the singular forms "a", "an", and "the" include plural references unless the context clearly dictates otherwise. Although any systems, methods, apparatuses, and devices similar or equivalent to those described herein can be used in the practice or testing of embodiments of the present invention, the preferred, systems and parts are now described. In the following description for the purpose of explanation and understanding reference has been made to numerous embodiments for which the intent is not to limit the scope of the invention.
One or more components of the invention are described as module for the understanding of the specification. For example, a module may include self-contained component in a hardware circuit comprising of logical gate, semiconductor device, integrated circuits or any other discrete component. The module may also be a part of any software programme executed by any hardware entity for example processor. The implementation of module as a software programme may include a set of logical instructions to be executed by the processor or any other hardware entity. Further a module may be incorporated with the set of instructions or a programme by means of an interface.
The disclosed embodiments are merely exemplary of the invention, which may be embodied in various forms.
The present invention discloses a system and method facilitating an entity selection over a communication network. A user may first access one or more catalogue through his communication device while communicating with an Interactive system. The catalogue provides him a list of items. The user may enter his choice in numerical form. This numerical form is a representation of the entity's name which is then processed by the interactive system to construct a suitable entity name by mapping and obtaining closest results from the pre-defined entity names which are stored in one or more database with respect to each numeric combination. The system and method also facilitates filtration of results by sorting out irrelevant results with respect to the user's numeric selection.

In accordance with an embodiment, referring to figure 1, the system (100) comprises of a communication device (102) communicating with an interactive system (104) for the entity selection over the interactive communication network. The interactive system (104) further comprises of a location identification module (106), an entity mapping engine (108) and a result filtration module (110). The system (100) further comprises of an output generation module (114) to validate the user's selection for one or more utility purpose.
Still referring to figure 1, by way of a specific example, the system (100) facilitates the entity selection over the interactive communication network. One or more user may access one or more catalogue through his communication device (102), containing list of entities without any need of going to entity distribution unit. The user may also provide his selection in order to obtain the selected entity.
The communication device (102) may include but is not limited to cell phone, laptop, desktop etc. These communication devices (102) are provided with a keypad for entering the suitable option. Selecting an item by typing name may be time consuming. The system (100) provides an option of entity selection by pressing numbers from the keypad of the communication device (102) which is easier and faster than entering text.
The selection of the entity in the numerical form will be a representation of the entity name. This numerical selection will then be further mapped to the entity names stored in the database (103) of the system (100).
The system (100) makes use of basic building blocks of any spoken language i.e. phonemes. These phonemes are the representation of words in spoken language. Every digit from 0-9 has a set of phonemes and the pronunciation of these digits is different in every language.
As an exemplary embodiment, Table 1 shows digit with their pronunciation in four different languages. Let us assume that generally people think or write in their native language or English not but a different language. A person X speaks Hindi then X may

think in Hindi or English but not Tamil. They may be able to speak many languages, but they will not think or write different language to order a product e.g. Soap. We have categorized products according to language and some common English words. Here product refers to entity.

Table 1: Digits in four languages
The system (100) uses the digits 0-9 to represent phonemes of that language. Thus, users can check the phoneme present in the digits and the phonemes in the item name. For example, in Marathi phoneme we can use 4 to represent "r" (T), etc. This is how each item
can be entered without actual text entry. Item names would be constructed based on the item name phonemes and number key phonemes mapping.
Users may be literate or illiterate, but they definitely speak their native language. It is also observed that even illiterate people may not be able to read but they understand numbers. Thus, users can use spoken language and numbers present on the keypad to order any product. Users can intuitively map the phonemes of the item to the phonemes of the numbers on the keypad, and keys in the numbers to specify the item name.

The system (100) further comprises of the interactive system (104) which is supporting the entity selection and is further configured to process the numeric selection in order to retrieve one or more closes results by constructing the entity name with respect to the numeric selection performed by the user.
The interactive system (104) further comprises of a location identification module (106) configured to identify a location of the communication device so that the user can select the entity name with respect to his native language or the in a language which makes the selection easy for him. The interactive system (104) further comprises of switching module (112) which is configured to allow the user to switch from one language to another.
The system (100) selects the default language for the user based on geo-coded location/address. The system (100) may be deployed in say Maharashtra, but if it gets a call from Guwahati, the default language for the user will be Assamese. This system (100) can be extended globally on interoperation agreements between telecom service providers. But is the user can understand Marathi and wants to switch from Assamese to Marathi, the user can change the language by way of the switching module (112). Since, the number of languages may exceed than 9 (as the key pad has only numeric digits 0-9), the user will simple enter the language selection in numeric form only as he does for entity selection. Again, this numeric selection will represent the name of language. The system (100) will then construct the language by this numeric selection and will support the entity selection according to the selected language.
Let us understand the mechanism of entity/language selection in numeric form:
Example: Set of phonemes according to numbers
Following phonemes are in Marathi. These can also be applicable to different languages. If the word is not present in the phonemes of numbers then the alphabets close to each other are used. If the phoneme is not present in the number phonemes then the closest alphabet of that language is used and putted in the same set.

To find out the users' language from the specific geography of the call origin a Hidden Markov Model is used. Let us consider India has this kind of IVR based system. India has 22 official languages which means if a person calls the system the phonemes, they will use, will be among these languages only. Suppose the system is deployed in Maharashtra state. A user calls from Maharashtra itself. For some instances, user selects Marathi as a native language and enters item name in Marathi language only. Then the probability of Marathi language increases and the system (100) shows only result obtained from Marathi language database.
Thus, the user selects an item by just entering name of the items using phonemes present in the item name and phoneme present in the digit. If the phoneme is not present in the phonemes of digits 0-9 then we use the phonemes close to the alphabet present in the list.
For example, is present in 4, but is not present in any of the number phonemes. Thus, we use 4 for also.
Here the Hidden Markov Model is used to identify whether user has meant or Initially the results which include both and will be shown, but as the system constructs the

probability of people choosing over increases, the system will give more preference to over for that product.
The interactive system (104) further comprises of the entity mapping engine (108) configured to correlate the numerical selection with a set of pre-identified numerical selections for a particular entity. The pre-identified numerical selections which may be performed by the user are constructed for each entity listed in the catalogue.
For example, for the entity A, the user may enter one of the numerical selections 1100, 1010, 1001 and so on. All these permutation-combination are stored in the database of the system (100) with respect to the languages thus supported by the system (100).
The entity mapping engine (108) then maps the correlated numerical selections with the one or more corresponding predefined entity names thus stored in the database (103) for the user selected language.
In order to retrieve the closest and correct results, the interactive system (104) further comprises of the result filtration module (110) configured to apply a selection mechanism by way of an embedded algorithm in order to sort (remove) irrelevant results from the mapped entity names. The result filtration module (110) is further configured to select a predetermined number of closest results with respect to the user's entity selection. By way of a specific example, the interactive system (104) may retrieve 3 closest results with respect to the user's numerical selection. If the system (100) will find the desired entity for the entry by the user then the system (100) will use the exact match only. But, if the number of search result is say 3 then the system may ask user to choose the correct one.
Below is disclosed an embedded algorithm for entity selection:
Embedded Algorithm:
Step 1: Find(S) where S is the input string by the user

Find(S) this function searches each string present in the database. It calls the functions compare(), sort().
Step 2: CompareQ: this compares each string.
Comparison is performed character by character.
Each phoneme has a set of possible digits. If the digits belong to same string then the result is zero.
The comparison function returns integer which is the difference between two strings.
Thus, the compare function calculates the differences between digits and returns differences.
By way of specific example, if a user enters product name XYZ as 12345. The database has the entry as 12344 for the same product. The pre-mapped set says 3, 4 and 5 belongs to same set, than the difference between 12345 and 12344 becomes zero. As such even if a user enters 12333 or 12444 or 12555it means the same and the difference is zero (0).
The HMM model is also provided in the compare function. This helps the result filtration module (110) to identify the closest match for the given string. The HMM model should be there in Compare which will first get the probability counts and then calculate the differences. This will help the module to get more accurate result.
By way of an example, English speakers use 4 for entering product name which has P in it. The digit 4 has f-o-u-r phonemes in it. They use 4 because O and P are closer. As more people uses 4 for P the system (100) will have higher probability for P than O. Thus, even before calculating the difference P is used instead of O to find the product into the system (100).
Step 3: sort(): This function sorts the differences according to increasing order. This function has HMM modeled into it. The HMM function finds the probability of a phoneme

over other and gives result to the user. First three results are shown to the user by the system (100).
End
Sample code:
Enter n where n is string of digits corresponding to the item name
psl is a character pointer
s is a string
L is the language
pps is a pointer to character pointer
char** find(const char* psl)
{
char** pps;
for every string s in L
{
if(strlen(s) == strlen(psl))
{
*pps - compare(s, psl);
pps++; } }
sort(*pps); //this function sorts in ascending order and also uses HMM to obtain results
return pps;
}
int compare(const char* psl, const char* ps2)
{
int diff;
while((diff=checkNumSet(*psl,*ps2))==0)
{
psl++; ps2++;
}
return diff;
}

int checkNumSet(int cl, int c2)
{
If(cl,c2£same set)
return 0;
else
return cl-c2;
}
Once, the interactive system (104) retrieves one or more closest results with respect to the numerical selection of the user, the user may further finalize one of the closest results thus provided by the interactive system (104). If all the results that are retrieved by the interactive system (104) does not matches to the entity thus desired by the user, the user may again provide his numerical selection to start the method of entity selection again. The system (100) further comprises of the output generation module (114) configured to accept the user's entity selection from the closest mapped results. The output generation module (114) is further configured to validate the user's selection for one or more utility purpose. The output generation module (114) checks whether the product desired by the user is properly mapped or matched to the result retrieved by the system (100). One or more utility purpose may include but is not limited to shopping though IVR. After validating the user's entity selection, the output generation module (114) may connect the user to a payment gateway in order to successfully obtain the entity. It may also receive an order from the user with respect to the entity thus selected to be shipped/delivered later.
The present system (100) retrieves results faster as compared to the systems disclosed in
the prior arts.
The efficiency of the system (100) is described below by way of a specific example.
The selection of a particular entity using the system (100) disclosed in the present invention
will be at least 3-5 times efficient (in terms of time) than selection of the same item using
some other IVR system using traditional drill down approach for catalogue navigation and
item selection.
Traditional drill down approach:

In traditional approaches, the IVR system present multiple layers of the catalogue to the
user and user presses digits to drill down for selecting a menu item present at the node of
this tree. In each layer users get exposed to multiple items to choose and go to next level of
the catalogue. After examining various traditional IVR systems we could assume that,
length of these recordings at any level is typically around 20 seconds.
If a user needs to drill down n no. of layers to select any item,
So Total time taken to choose a particular item is :
n*(time_at_each_layer)+ n*(time_to_press_one_digit) = n(time_at_each_layer +
time_to_press_one_di git)
where n is number of items present in the list of that layer.
If users want to select same item using the proposed system, at the beginning of the system the user will specify encoding for their query. This is obtained by the location based service of the system.
In traditional approach, at every layer user selects 1 option from all the options. If he/she traverses n number of layers down the tree, this means he/she would have selected n different options.
In the present system, the user at the beginning of the system will enter encoding corresponding to these n options.
If we assume that every word has 8 characters, then it will have 8*n characters. We may assume that 1 phoneme consists of 2 characters then the encoding will have, (8*n)/2 phonemes = 4*n phonemes
In the present system each phoneme is represented by a digit. Hence time taken to enter 4*n phonemes = (4*n)*time_to_press_one digit.
This is nothing but total time taken to choose a particular item using our approach. We may assume that, timeateachlayer » time_to_press_one_digit
Let us call time_at_each_layer = A time_to_press_one_digit = B

And we have assumed that A » B
Now,
Total time taken to choose a particular item using traditional approach = (n*A + n*B)
Total time taken to choose a particular item using our approach = 4*n*B
Hence, the ratio, (time taken to choose a particular item using traditional approach) /(time
taken to choose a particular item using present invention) would turn around,
(n*A + n*B)/(4*n*B)
= (A+B)/4B
= 0.25*(A/B+1)
It is observed that A » B
We can even assume that, A ~ 20*B (Approx.)
Considering this relation we can observe that, the above ratio would turn out to be
= 0.25*(20*B/B + 1)
~ 5 (Approx.)
i.e. (time taken to choose a particular item using traditional approach) /(time taken to
choose a particular item using our approach) = 5
Therefore, time taken to choose a particular item using traditional approach ~ 5*Time
taken to choose a particular item using our approach.
BEST MODE/EXAMPLE FOR WORKING OF THE INVENTION
The system and method illustrated to facilitate an entity selection in an interactive communication network may be illustrated by working example stated in the following paragraph; the process is not restricted to the said example only:
Referring to figure 2 and 3, Let a user calls in an IVR number to purchase an item (entity). The IVR system plays a welcome message (step 302). System then prompts for product name (step 304)

User then Keys in the item to search the catalogues (step 306)
In this step, the IVR prompts the user to enter the item query. For a first time user, the text-entry mechanism is explained. For frequent users, the mechanism is available as an option upon a specific (help) key press.
The text entry method is phoneme based and includes the following: User keys in a query string by closely mapping the phoneme sounds of the phrase, to phoneme sounds of the numbers 0-9.
The system is designed in a manner, where each item presented in an IVR menu would be given an encoding in the form of numbers based on following approach. Choose a language
Split every digit from 0 to 9 into its phonetics. E.g. Hindi Language
0 shunya - sha na u na ya a
1 ek - ae ka
2 do - da o
3 teen - ta ee na
4 chaar - cha aa ra
5 paanch - pa a na cha
6 chhe - chha ae
7 saat - sa a ta
8 aath - aa tha
9 nou - na o u
Suppose user wants to purchase "Dhaniya"
The system will ask user to encode this word using logic mentioned above
A typical encoding for this word would be
Dha ~ da 2
na~ na 9, 3, 5
i ~ee 3
ya ~ ya 0

aa ~ aa 4, 5, 7, 8
This possible mapping from user may lead to entry of one of the following 12 sequences to
be keyed in by the user
29304
29305
29308
29307
23304 23305 23308 23307
25304 25305 25308 25307
One user will at a time enter one of the above sequences (Not all)
Say he has entered 23304.
Now system (entity mapping engine of the interactive system) will try to find closest match
for this string (step 308) by checking if there is a match for string of digits (step 310).
If there is no match system will get back to step 304.

Item Name Hindi Phonemes
Shirt 0447
Pants 58077
Dal 247
Dhaniya 23304
Table 2 If a match is found in table 2, say system has stored "Dhaniya" in Hindi as 29304. Say there are other words whose encoding is closer to 23304.
Say there are words corresponding to digits 21304, 22304, 24304, 26304, 27304, 28304. (Then we will choose 29304 out of these as a closest map to 23304 because digits 3, 9 fall under same phonetics (na) category for Hindi language. See phonetics distribution of every digit mentioned above.)
When there is more than one match which is close to a user's selection, system will plays at max 3 matches for the user (step 312). User will then select an item from the matched choice (step 314).
Therefore, the user will obtain the correct product matching to his selection. Upon successful validation this product may be ordered and later may be shipped to user at his desired location (step 316 and 318).

WE CLAIM:
1. A system facilitating an entity selection over an interactive communication network, the system comprising:
a communication device configured to provide an access to a user to one or more catalogue containing a list of entities, the communication device upon successful access further configured to select an entity in a numerical form such that the numerical form is a representation of the entity name;
an interactive voice system communicating with the communication device, configured to process the numerical selection in order to retrieve one or more closest results by constructing the entity name with respect to the numerical entity selection, the interactive voice system comprising:
a location identification module configured to identify a location of the
communication device in order to select the language for the catalogue thus
accessed by the user;
an entity mapping engine configured to correlate the numerical selection
with a set of pre-identified numerical selections for a particular entity and
map the correlated numerical selections with the one or more corresponding
predefined entity names thus stored in a database;
a result filtration module configured to apply a selection mechanism to sort
irrelevant results from the mapped entity results and select a predetermined
number of closest results with respect to the user's entity selection; and
an output generation module configured to accept the user's entity selection from the closest mapped results and validate the user's selection for one or more utility purpose.
2. The system as claimed in claim 1, wherein the entity further comprises of one or more products to be shopped and are listed in the catalogue.

3. The system as claimed in claim 1, wherein the communication device may include but is not limited to a mobile phone, a laptop, a desktop.
4. The system as claimed in claim 1, wherein the numerical form of the entity thus selected further comprises a combination of numerical digits from 0 to 9.
5. The system as claimed in claim 1, wherein the interactive communication system retrieves 3 closest mapped results with respect to a numerical selection.
6. The system as claimed in claim 1, wherein the system further comprises of a switching module configured to allow the user to switch from one language to another.
7. The system as claimed in claim 1, wherein the utility purpose may include but is not limited to shopping by using interactive voice system.
8. The system as claimed in claim 1, wherein the database stores the predefined entity names in a plurality of languages.
9. A method facilitating an entity selection over an interactive communication network, the method comprising steps of:
providing an access to a user to one or more catalogue containing list of entities to and selecting the entity in a numerical form, such that numerical form is a representation of the entity name;
processing the numerical selection in order to retrieve one or more closest results by constructing the entity name with respect to the numerical entity selection, the processing further comprising:
identifying a location of the user in communication in order to select the
language for the catalogue thus accessed by the user;

correlating the numerical selection with a set of pre-identified numerical
selections for a particular entity and map the correlated numerical selections
with the one or more corresponding predefined entity names thus stored in a
database;
applying a selection mechanism to sort the irrelevant results from the
mapped entity results and select a predetermined number of closest results
with respect to the user's entity selection; and
accepting the user's entity selection from the closest mapped results and validate the user's the user's selection for one or more utility purpose.
10. The method as claimed in claim 9, wherein the entity further comprises of one or more products to be shopped and are listed in the catalogue.
11. The method as claimed in claim 9, wherein the numerical form of the entity thus selected further comprises a combination of numerical digits from 0 to 9.
12. The method as claimed in claim 9, wherein the method further retrieves 3 closest mapped results with respect to a numerical selection.
13. The method as claimed in claim 9, wherein the method further comprises of step of allowing the user to switch from one language to another.
14. The method as claimed in claim 9, wherein the utility purpose may include but is not limited to shopping by using interactive voice system.
15. The method as claimed in claim 9, wherein the method further comprises of storing the predefined entity names in a plurality of languages.

16. The method as claimed in claim 9, wherein the correlation of the numerical selection is further performed by using a Hidden Markov Model.