FORM 2
THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENT RULES, 2003
COMPLETE SPECIFICATION
(See Section 10 and Rule 13)
Title of invention:
A METHOD AND SYSTEM FOR DEMAND FORECASTING FOR SEASONAL PRODUCTS FOR PRICING DURING OFF-SEASON
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 THE INVENTION
The invention generally relates to retail price optimization for products. Particularly the invention relates to a method and system for demand forecasting for seasonal products for pricing during off-season.
BACKGROUND OF THE INVENTION
Seasonality or local market environment plays an important role on product sales. Most of the products sales pattern follows the seasonal trend, i.e., it tends to peak at certain time point of the year and diminishes very fast after peak time and eventually after three to four weeks, demand tends to come down to zero. The demand before peak-period is known as in-season demand period and after peak-period is known as off-season demand period. On an average the sales during offseason period is one-fourth to one-fifth of complete year sales. In general, the customer purchases the product during off-season because of low price offering and plan to use the product for next year's seasonal event. On an average a store is left with one-tenth to one-twentieth of their total inventory because of inappropriate price offering and conflict of customer's interest. Thus, accurate demand forecasting of a seasonal product during off-season may improve store efficiency and revenue, significantly.
Normally, retailers or distributors calculate the demand of the product in market based on historical data such as billings or sales from the retailers or distributers or from stores of sales. However, in case of seasonal product, the products attribute are changing dynamic, year on year and same product might not appear in subsequent years. Hence, the historical data availability may not directly map the future product sales and always tends to imprecise demand forecasting for seasonal product. Moreover, the off-season demand behaviour may not match or sink with in-season demand behaviour. In addition, the price discount needs to be offered for a group of products for off-season instead of a single product and each product might belong to a different class of products.

Most of the conventional methods for demand forecasting for a product (or set of products) for pricing are based on historical transaction data of the same product (or set of products). Thus, there is a long standing need to improve the efficiency in the process for forecasting demand of a seasonal product for pricing during offseason, when the past historical transaction data is not available and only current year complete or partial in-season data is available. There is also a need to provide a method and system which is capable of creating a plurality of virtual products and corresponding transaction data, and further estimating price-elasticity for forecasting demand based upon that.
OBJECTIVES OF THE INVENTION
The primary objective of the present invention is to provide a method and system for demand forecasting for seasonal products for pricing during off-season in absence of the past transaction data and the availability of complete or partial current year in-season data.
Another objective of the invention is to provide a method and system for creating a plurality of virtual products and corresponding transaction data based on similarity matrix & clustering algorithms.
Another objective of the invention is to provide a method and system for normalizing the plurality of virtual products and corresponding transaction data of virtual products at the same price-elasticity demographic level as current seasonal product for representing past year transactional data. Further, a method and system is provided for normalizing and rescaling the current year in-season data to make it same level with virtual product.
Another objective of the invention is to provide a method and system for estimating the impact of different factors at different hierarchy levels such as product level, demographic level based on price-elasticity behaviour, demographic level based on weather for representing non-linear behaviour of

price-elasticity model or sales response function for off-season period on the plurality of virtual products and current seasonal products at different hierarchy level.
Another objective of the invention is to provide a method and system for determining the weightage needed to be considered on past year forecast errors on the plurality of virtual products and corresponding transaction data for further deriving/updating the impact of each of the factors on sale during current year offseason period using weightage Bayesian method.
SUMMARY OF THE INVENTION
Before the present methods, systems, and hardware enablement are described, it is to be understood that this invention in not limited to the particular systems, and methodologies described, as there can be multiple possible embodiments of the present invention which are not expressly illustrated in the present disclosure. It is also to be understood that the terminology used in the description is for the purpose of describing the particular versions or embodiments only, and is not intended to limit the scope of the present invention.
The present invention provides a method and system for demand forecasting of seasonal products for pricing during off-season period in absence of the past year transaction data and the presence of current year complete or partial in-season data.
In an embodiment, a method is provided for demand forecasting of seasonal products for pricing during off-season period, characterized by demand forecasting in absence of past transaction data and presence of current year complete or partial in-season data. The method comprises processor implemented steps of creating a plurality of virtual products and corresponding transaction data based on current and historical product attributes and current year in-season and historical transaction data using a virtual product and transaction data creation

module; normalizing the plurality of virtual products, corresponding transaction and the current year in-season transaction data and further re-scaling normalized current year in-season transaction data as per peak-period and peak-sales of virtual product using a virtual product transaction and current year in-season data normalization module; estimating the impact of at least one factor on the plurality of virtual products and current season products at different hierarchy level for representing non-linear behaviour of sales response function for off-season period using a sales response function design and estimation module; determining the weightage need to be considered for the plurality of virtual product and corresponding transaction data of past years for further deriving or updating the impact of each the factor on off-season period using a sales response function updation module; and forecasting demand by utilizing different hierarchies of the attributes using a demand forecasting module.
In an embodiment of the invention a system is provided for demand forecasting of seasonal products for pricing during off-season period, characterized by demand forecasting in absence of the past transaction data and the presence of current year complete or partial in-season data. The system comprises of a virtual product and transaction data creation module adapted to create a plurality of virtual products and corresponding transaction data based on current and historical product attributes and current year in-season and historical transaction data; a virtual product transaction and current year in-season data normalization module adapted to normalize the plurality of virtual products, corresponding transaction and the current year in-season transaction data and further re-scaling normalized current year in-season transaction data as per peak-period and peak-sales of virtual product; a sales response function design and estimation module adapted to estimate the impact of at least one factor on the plurality of virtual products and current season products at different hierarchy level for representing non-linear behaviour of sales response function for off-season period; a sales response function updation module adapted to determining the weightage need to be considered for the plurality of virtual product and corresponding transaction data of past years for further deriving or updating the impact of each the factor on off-

season period; and a demand forecasting module adapted to forecast demand by utilizing different hierarchies of the attributes, wherein the virtual product and transaction data creation module; the virtual product transaction and current year in-season data normalization module; the sales response function design and estimation module; the sales response function updation module; and the demand forecasting module are coupled to an electronic hardware processor configured to perform operative functions of each.
The above the method and system are preferably for demand forecasting of seasonal product for pricing during off-season period in absence of the past transaction data and the presence of complete or partial current year in-season data, but also may be used for many other applications as may be clear to a person skilled in the field of demand forecasting.
BRIEF DESCRIPTION OF DRAWINGS
The foregoing summary, as well as the following detailed description of 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 example constructions of the invention; however, the invention is not limited to the specific methods and apparatus disclosed in the drawings:
Figure 1 shows a flowchart (100) illustrating the method for demand forecasting of seasonal product for pricing during off-season period.
Figure 2 shows a flowchart (200) illustrating the method for creating virtual product and transaction data.
Figure 3 shows a flowchart (300) illustrating the method for estimating parameters of price elasticity/sales response function.

Figure 4 shows a flowchart (400) illustrating the method for calculating impact of climate with sales units based on a linear regression technique.
Figure 5 show s a flowchart (500) 11 lustrating the method for price-elasticity forecasting model using Bayesian method.
Figure 6 shows a flowchart (600) illustrating the method for forecasting at store group level.
Figure 7 shows a block diagram illustrating the system (700) for demand forecasting of seasonal product for pricing during off-season period.
Figure 8 shows a block diagram illustrating the system architecture (800) for demand forecasting of seasonal product for pricing during off-season period.
DETAILED DESCRIPTION OF THE INVENTION
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, 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. The

disclosed embodiments are merely exemplary of the invention, which may be embodied in various forms.
The disclosed embodiments are merely exemplary of the invention, which may be embodied in various forms.
The present application provides a method for demand forecasting of seasonal products for pricing during off-season period, characterized by demand forecasting in absence of past transaction data and presence of current year complete or partial in-season data, the method comprises processor implemented steps of:
a. creating a plurality of virtual products and corresponding transaction
data based on current and historical product attributes and current year
in-season and historical transaction data using a virtual product and
transaction data creation module (702);
b. normalizing the plurality of virtual products, corresponding transaction
and the current year in-season transaction data and further re-scaling
normalized current year in-season transaction data as per peak-period
and peak-sales of virtual product using a virtual product transaction and
current year in-season data normalization module (704);
c. estimating the impact of at least one factor on the plurality of virtual
products and current season products at different hierarchy level for
representing non-linear behaviour of sales response function for off
season period using a sales response function design and estimation
module (706);
d. determining the weightage need to be considered for the plurality of
virtual product and corresponding transaction data of past years for
further deriving or updating the impact of each the factor on off-season
period using a sales response function updation module (708); and
e. forecasting demand by utilizing different hierarchies of the attributes
using a demand forecasting module (710).

The present application also provides a system for demand forecasting of seasonal products for pricing during off-season period,' characterized by demand forecasting in absence of the past transaction data and the presence of current year complete or partial in-season data, the system comprises of:
a. a virtual product and transaction data creation module (702) adapted to
create a plurality of virtual products and corresponding transaction data
based on current and historical product attributes and current year in-
season and historical transaction data;
b. a virtual product transaction and current year in-season data
normalization module (704) adapted to normalize the plurality of virtual
products, corresponding transaction and the current year in-season
transaction data and further re-scaling normalized current year in-
season transaction data as per peak-period and peak-sales of virtual
product;
c. a sales response function design and estimation module (706) adapted to
estimate the impact of at least one factor on the plurality of virtual
products and current season products at different hierarchy level for
representing non-linear behaviour of sales response function for off
season period;
d. a sales response function updation module (708) adapted to determining
the weightage need to be considered for the plurality of virtual product
and corresponding transaction data of past years for further deriving or
updating the impact of each the factor on off-season period; and
e. a demand forecasting module (710) adapted to forecast demand by
utilizing different hierarchies of the attributes, wherein the virtual
product and transaction data creation module (702); the virtual product
transaction and current year in-season data normalization module (704);
the sales response function design and estimation module (706); the
sales response function updation module (708); and the demand
forecasting module (710) are coupled to an electronic hardware
processor configured to perform operative functions of each.

Referring now to Figure 1, a flowchart (100) illustrating the method for demand forecasting of seasonal product for pricing during off-season period is described.
The process starts at step 102, the virtual product and corresponding transaction data is created to represent the current seasonal products and derive the past years transactional data using similarity matrix and clustering algorithm. Further, at step 104, the created virtual product transaction data and current year in-season transaction data is normalized to equivalent format. The current year in-season normalized data is rescaled to align the impact of peak-period or peak-sales. At step 106, the price-elasticity/sales response function is designed. At step 108, the impact of different factors is updated for each week during off-season based on forecast errors on virtual product sales at past years using weighted Bayesian method. The process ends at step 110, where demand is forecasted at store and daily level using forecasting model.
Figure 2 is a flowchart (200) illustrating the method for creating virtual product and transaction data.
The process starts at step 202, the product for off-season pricing is identified. At step 204, the current product attributes are compared with historical product attributes and transaction data. At step 206, similar products are found using clustering algorithms and are named as virtual product. At step 208, the stores are clustered based on price elasticity parameters such as discount percentage, sales increment percentage and peak sales indicators which are generated from transaction data, to represent store group. At step 210, the transaction data related to virtual product is extracted. At step 212, the extracted transaction data related to virtual product is transformed into similar data as the current year in-season products data. At step 214, the transformed transaction data is normalized to make the virtual product's transaction data to the same level as that of the normalized current year in-season product's transaction data. The process ends at step 216, the current product group transaction data is normalized and rescaled based on projected peak period and peak sales.

Referring to Figure 3, that is a flowchart (300) illustrating the method for estimating parameters of price elasticity/sales response function.
The process starts at step 302, the price elasticity model estimates using the factors such as product price discount, peak-period deviation, product advertisement, and at the store group level store grouping is based on same price elasticity and peak sales impact using least square method. At step 304, the inventory parameters at store level are updated using Bayesian method. The process ends at step 306 where the impact of climate with sales units is calculated based on a linear regression technique.
Referring to Figure 4, that is a flowchart (400) illustrating the method for calculating impact of climate with sales units based on a linear regression technique.
The process starts at step 402, the stores are clustered based on same weather conditions at a particular time interval such as monthly or weekly bucket level. At step 404, the sales units are standardized using maximum and minimum scaling approach. At step 406, the deviation of different store group from the base climate is found. The process ends at step 408 where the impact of climate with sales units is calculated based on a linear regression technique.
Referring to Figure 5, that is a flowchart (500) illustrating the method for price-elasticity forecasting model using Bayesian method.
The process starts at step 502, the forecast model and parameters error of initial price elasticity model are calculated and considered as initial parameters for Bayes algorithm. At step 504, the sales data for past years for 1st week of off-season is forecasted by using the initial price elasticity model parameters. At step 506, errors in 1st week of off-season for past years are calculated. At step 508, weightage is given for past year data based on weightage method (e.g. regression) to represent the combined error. At step 510, the price elasticity parameters are

updated by utilizing combined error by using Bayesian method. At step 512, the forecast model and coefficient error is updated based on Bayesian method. The process ends at step 514 where the parameters for next week of off-season period are updated by considering the updated parameters as initial parameters.
Referring to Figure 6, that is a flowchart (600) illustrating the method for forecasting at store group level.
The process starts at step 602, forecast at store group level for different week using the Bayesian model. The Bayesian model selection may use the rules of probability theory to select among different hypotheses. At step 604, the forecast sales are distributed into different store levels. At step 606, the sale units are forecasted at store level for each week using inventory correction. At step 608, the correction for climate factor is forecasted for a given week is done. The process ends at step 610 where the weekly sales are converted into daily sales.
Figure 7 is a block diagram illustrating the system (700) for demand forecasting of seasonal product for pricing during off-season period.
The system (700) for demand forecasting of seasonal product for pricing during off-season period comprises of a virtual product and transaction data creation module (702); a virtual product transaction and current year in-season data normalization module (704); a sales response function design and estimation module (706); a sales response function updation module (708) and a demand forecasting module (710).
In an embodiment of the present invention, the virtual product and transaction data creation module (702) is adapted to create a plurality of virtual products for representing current seasonal products and corresponding transaction data based on current product attributes and transaction data behaviour. The plurality of virtual products is created based on current product attributes and historical product category comparison by finding the most similar products using similarity

matrix and clustering method. The seasonal product may comprise of any product produced by the manufacturing company for sale or promotion of the product. The transactional data may comprise of all the information related to the product which may further comprise of the information like total product sale at particular time, profit retrieved from the product, demand of the product, queries related to the product and the like.
In an embodiment of the present invention, the virtual product transaction and current year in-season data normalization module (704) is adapted to normalize the plurality of virtual products and corresponding transaction data. The plurality of virtual products and corresponding transaction data are normalized at a same price-elasticity demographic level as current seasonal product for representing past year transactional data.
The normalization of data may be the division of multiple sets of data by a common variable in order to nullify that variable's effect on the data, thus allowing underlying characteristics of the data sets to be compared which allows data on different scales to be compared, by bringing them to a common scale or current seasonal data.
In a working example of virtual product creation and normalization, a business user may select current year products A, B, and C to decide off-season pricing for
the same. Considering the total number of seasons, {1 ...., tp,..,tpeak, ,T}.
Further, only current partial in-season {1 ...,,tp} data are available for products A, B, and C for forecasting demand for pricing of the same. Considering that the Peak-sales happened in-season at tpeak, the off-season shall be considered as
{tpeak+i, ,T}. There is a need to forecast for the product A, B, C for off-season
period for different discount or pricing where the historical transaction data of products A, B, and C are not available. Considering the historical similar class (based on product hierarchy) of products A, B and C as follows:
A→{A1,A2, A3}

B→{B1,B2, B3} C→{C1,C2,C3}
For virtual product selection for each current product, consider product A and the attributes for the product A may be K, (quality), K2 (price), K3 (weight),....,Kn Demand pattern parameters of product A may be considered as M1 (peak-period), M2 (% increase in demand due to change in % price), coefficient of variance (M3)...., Mq. Also, the same attributes and demand pattern parameters of Al, A2, and A3 can be derived from historical transaction data. Table 1 below describes the attributes and demand pattern parameters of product A and its similar class products:

Ki K2 K3 M, M2
A High 10 5 22 1
Al Med 8 2 20 1.3
A2 High 14 7 25 1.5
A3 Low 6 4 18 1
Table 1: Attributes and demand pattern parameters of product A and its similar
class products
The similarity matrix is created using different techniques, such as Gower's distance technique is used when the attributes and demand pattern parameters are nominal, ordinal and ratio variables type. Table 2 below describes the similarity matrix between products A, Al, A2, A3.

A Al A2 A3
A 1 0.8 0.7 0.6
Al 0.8 1 0.6 0.4
A2 0.7 0.6 1 0.7
A3 0.6 0.4 0.7 1

Table 2: Similarity matrix between products A, Al, A2, A3
Based on similar matrix given in Table 2, the clustering technique is used to the most similar classes of product A. Considering that Al is the most similar class of product A from the above analysis. Similarly, consider that, Bl and CI classes are the most similar class of products, B and C, respectively. Hence, the virtual product of A, B, C, can be represented as combination of Al, Bl, and CI.One of the ways that can be used for representation of the virtual products transaction data, such as the transformation of historical sales unit data of virtual product is given below:
Consider, the average sales units of product A, B, and C are SalesA, SalesB, and SalesC, respectively.
The contribution of products A, B, and C of total sales are RA = Sales_A/(Sales_A + SalesB + Sales_C), RA = Sales_B/(Sales_A + Sales_B + Sales_C), and Rc = Sales_C/(Sales_A + Sales_B + Sales_C).
Consider the normalization of the sales units of Al, Bl, and CI using peak sales of Al, Bl, and CI, respectively, and the normalized values are NA1, NB1, and Nc1 respectively. Then the virtual products transaction data of sales units are represented by:
NA1*RA + NB1*RB + Na*Rc.
Similarly, the other transaction data such as sales dollar, inventory, and discount of virtual products are represented. In this way, the transaction data of virtual product are represented for past three years for whole season (1,.. .,tpeak,......,T).
Considering the normalization of the current year transaction data for products (A, B, and C), the current year data is normalized using combined peak sales and is represented as current year combined normalized data. The peak sales time period is also known for virtual product. For reseating of the current year transaction

data, the current year peak sales period is based on partial current year information. Hence, there is a need to rescale the normalized current year transaction data. Considering the actual peak period of current year products are same as the virtual product's peak period, the extrapolation method is used to adjust the normalized current products normalized data.
In an embodiment of the present invention, the sales response function design and estimation module (706) is adapted to estimate the impact of different factors on the plurality of virtual products and current season products at different hierarchy level for representing non-linear behaviour of sales response function for offseason period, where in the hierarchy used to design the base price-elasticity and forecasting model is selected from the group comprising of product, demographic based on price, and demographic based on weather. The factors may be a climatic factor, market condition factor, and the demand factor such as pricing of the product, life-cycle of the product, peak sales of products, peak period, length of partial in-season period, and linear model to represent the non-linear behaviour of sales response function for off-season period. The sales response function design and estimation module is further adapted to find peak sales, peak-period, length of partial in-season period.
The sales response function or parameters are estimated by following hierarchical steps: estimating price-elasticity parameters at store group level based on same price-elasticity impact using least square method; updating inventory parameter at store level using Bayesian method ; estimating the impact of climate by store grouping based on same climate behaviour by clustering the store at different point of time based on same weather; standardizing the sales units using maximum and minimum scaling approach; considering the base climate and find the deviation of different store group from the base climate and deviation of sales units for different store group; and using linear regression technique to find the impact of climate with sales units.

In an embodiment of the present invention, the sales response function updation module (708) is adapted to determine the weightage need to be considered for the plurality of virtual product and corresponding transaction data of past years for further deriving the impact of each the factor on off-season period, which is determined by weightage Bayesian method. The method comprises of updating the impact of price-discount and other factors on the plurality of virtual products and current season products for sale during off-season. The major steps are as follows: The parameters and their errors are estimated from base price elasticity model and consider the calculated base price-elasticity model and parameters errors as initial parameters for Bayes algorithm; forecasting the sales for past years for first week of off-season by using the base price elasticity model parameters; calculating error in 1st week of off-season of past years forecast giving weightage for past year data to represent the combined error by utilizing weightage method; updating the price elasticity parameters by utilizing combined error by using Bayesian method; updating the forecast model and coefficient error based on Bayesian method; and updating the parameters for next week of offseason period by considering the updated parameters as initial parameters.
In an embodiment of the present invention, the demand forecasting module (710), forecast the demand at different hierarchy level. The forecasting at store group level is done for different week using Bayesian model; distributing forecast sales into different store level; forecasting sales units at store level for each week using inventory correction; forecasting correction for climate factor for a given week; and converting weekly sales into daily sales.
Figure 8 is a block diagram illustrating the system architecture (800) for demand forecasting of seasonal product for pricing during off-season period.
The system architecture (800) for demand forecasting of seasonal product for pricing during off-season period comprises of a virtual bucket & clustering module (802); a store group estimation module (804); a sales response function design and estimation module and a sales response function updation module (706

& 708); a demand forecasting module (710); a virtual bucket database (806); a store group database (808); a price elasticity database (810); a demand forecasting database (812); a transactional database (814); an inventory database (816); a store climate information database (818); a promotion information database (820) and a network (822).
In an embodiment of the present invention, the virtual product and clustering module (802) is adapted to cluster the similar products with similar factors based on their characteristics, together inside a similar virtual bucket and stored in the virtual bucket database (806). The similar products are grouped inside a similar bucket using a clustering algorithm which may be further utilized to form a group of stores.
In an embodiment of the present invention, the store group estimation module (804) is adapted to intake all the information of different stores from the virtual bucket database (806) and the transactional database (814).
In an embodiment of the present invention the sales response function model comprises of an Subscript description: D[T, P, IR, C, A ], it represents demand of a Stock Keeping Unit (SKU) Group for a given time period (T), price discount (P), effective inventory (1), climate (C) and advertisement type (A). The SKU is a number or code used to identify each unique product or item for sale in a store or other business. T represents the lifecycles path/time of SKU bucket and TO indicates peak sales time period determine by current year data or past year data using a defined method, such as, linear interpolation / extrapolation. Fraction of deviation of lifecycle from positive (in-season) and negative (off-season) represented by (T-T0)+/N+ and (T -T0)-/N-)], respe ctively, wh ere N+ and redetermine based on number of observations in in-season and off-season and observation patterns. IR represents the % Inventory level, where IR =Max(I, It). I and It are current inventory level and effective threshold, respectively. These parameters are determined by categorical inventory level analysis and sales

patterns. Climate (C): It represents the reduction/increase on transaction due to climate change.
Example of a type of price elasticity estimation model structure is given below: Linear Model:
D[T, P, A, I, C] = al + a2 * T + a3 * [1 /(T +1)] * P + a4 * IR + a5*C + a6*A
in an embodiment of the present invention, the sales response function design and estimation module (706) is adapted to estimate the quantity of a product or service that consumers will purchase in future. Demand forecasting may be used in making pricing decisions, in assessing future capacity requirements, or in making decisions on whether to enter a new market. Demand forecasting in off-season may be decided based on the inputs given by the various modules in the system.
In an embodiment of the present invention, the store group database (808) may store the information related to the similar stores. The data stored may be in grouped format or may be separated format depending on the similarities between the stores.
In an embodiment of the present invention, the price elasticity database (810) is adapted to store all the information of the store group estimation module (804), the sales response function design and estimation module (706) and the sales response function updation module (708). The price elasticity database may store the future information of the quantity demanded in response to a one percent change in price of the product.
In an embodiment of the present invention, the demand forecast database (812) is adapted to store all the information of the demand forecasting module (710). The demand forecast database may be used to store the information quantity of a product or service that consumers will purchase in future. The demand forecast

database may be used find the impact related to the product like the climate effects, market response effects and the like.
In an embodiment of the present invention, the transactional database (814) is adapted to store all the transactional data related to the products and the stores. The transactional database may include the profit information related to the product, the sales requirement in near future, information related to the orders of the product, all the product activity records, and the like.
In an embodiment of the present invention, the inventory database (816) is adapted for storing all the information related to tracking of product stock, suppliers, employees, purchase orders, sales with this inventory management database and the like.
In an embodiment of the present invention, the store climate information database (818) is adapted to store all the information related to the climate or environmental conditions like effect of the climate on the product, the sale of product in different seasons with varying climate and the like.
In an embodiment of the present invention, the promotion information database (820) is adapted to store all the information related to the promotion of the product. The promotion information database may include the information related personal selling, advertising, sales promotion, direct marketing, publicity strategy and the like.
In an embodiment of the present invention, the network (822) is adapted to communicate with all the components of the system architecture (800). The network may be utilized for the inter-communication between the various components of the system architecture (800). The data stored in each component may be transferred between the various components of the system architecture (800).

WE CLAIM:
1. A method for demand forecasting of seasonal products for pricing during off-season period, characterized by demand forecasting in absence of past transaction data and presence of current year complete or partial in-season data, the method comprises processor implemented steps of:
a. creating a plurality of virtual products and corresponding
transaction data based on current and historical product attributes
and current year in-season and historical transaction data using a
virtual product and transaction data creation module (702);
b. normalizing the plurality of virtual products, corresponding
transaction and the current year in-season transaction data and
further re-scaiing normalized current year in-season transaction
data as per peak-period and peak-sales of virtual product using a
virtual product transaction and current year in-season data
normalization module (704);
c. estimating the impact of at least one factor on the plurality of
virtual products and current season products at different hierarchy
level for representing non-linear behaviour of sales response
function for off-season period using a sales response function
design and estimation module (706);
d. determining the weightage need to be considered for the plurality
of virtual product and corresponding transaction data of past years
for further deriving or updating the impact of each the factor on
off-season period using a sales response function updation module
(708); and
e. forecasting demand by utilizing different hierarchies of the
attributes using a demand forecasting module (710).

2. The method as claimed in claim 1, wherein the plurality of virtual products is created based on current and historical product attributes and the historical and current year in-season transaction data by finding the most similar products using similarity matrix and clustering method.
3. The method as claimed in claim 2, further comprises of extracting the current year product group sales and discount and clustering the stores based on price elasticity parameters to represent store group.
4. The method as claimed in claim 1, wherein the plurality of virtual products, corresponding transaction and the current year in-season transaction data are normalized and further re-scaled at a same price-elasticity demographic level as current seasonal product for representing past year transactional data and to adjust peak-period sales.
5. The method as claimed in claim 1, wherein the price-elasticity parameters or sales response function are estimated by processor implemented steps of:
a. estimating price-elasticity parameters at store group level based on
same price elasticity impact using least square method;
b. updating inventory parameter at store level using Bayesian method;
and
c. estimating the impact of climate by store grouping based on same
climate behaviour by clustering the store at different point of time
based on same weather; standardizing the sales units using
maximum and minimum scaling approach; considering the base
climate and find the deviation of different store group from the
base climate and deviation of sales units for different store group;
and using linear regression technique to find the impact of climate
with sales units.

6. The method as claimed in claim 5, wherein the sales response function or factors are selected from the group comprising of product price discount, peak-period deviation, product advertisement, and inventory parameter.
7. The method as claimed in claim 1, wherein the sales response function updation is done by processor implemented steps of:
a. calculating the forecast model and parameters error of initial price
elasticity model and consider the calculated forecast model and
parameters error as initial parameters for Bayes algorithm;
b. forecasting the sales for past years for first week of off-season by
using the initial price elasticity model parameters;
c. calculating errors for 1st week of off-season forecast for past years;
d. utilizing weightage method for giving weightage for past year data
to represent the combined error;
e. updating the price elasticity parameters by utilizing combined error
by using Bayesian method;
f. updating the forecast model and coefficient error based on
Bayesian method; and
g. updating the parameters for next week of off-season period by
considering the updated parameters as initial parameters.
8. The method as claimed in claim 7, wherein the forecasting at store group level is done by processor implemented steps of forecasting at store group level for different week using Bayesian model; distributing forecast sales into different store level; forecasting sales units at store level for each week using inventory correction; forecasting correction for climate factor for a given week; and converting weekly sales into daily sales.
9. The method as claimed in claim 1, wherein the factors are selected from the group comprising of climatic factor, market condition factor, and demand factor such as pricing of the product, life-cycle of the product,

peak sales of products, peak period, length of partial in-season period, and linear model to represent the non-linear behaviour of sales response function for off-season period.
10. The method as claimed in claim 1, wherein the different hierarchy is used to estimate impact of different factors selected from the group comprising of price-discount based on price-behaviour demographic, climate impact based on similar climate of different stores.
11. The method as claimed in claim 1, further comprises of clustering the similar virtual products based on current season products attributes and transaction data and stored in virtual bucket database. The virtual bucket database and transactional database are utilized to form a group of stores using a store group estimation module (804).
12. The method as claimed in claim 1, further comprises of clustering similar products with similar factors based on their characteristics, together inside a similar virtual bucket and storing the same in a virtual bucket database (806) using a virtual bucket & clustering module (802); inputting information pertaining to different stores from the virtual bucket database (806) and a transactional database (814) using a store group estimation module (804); estimating the quantity of a product or service that consumers will purchase in future using the sales response function design and estimation module (706); storing information pertaining to the similar stores using a store group database (808); storing information pertaining to the store group estimation module (804), the sales response function design and estimation module (706) and the sales response function updation module (708) using a price elasticity database (810); storing information pertaining to the demand forecasting module (710) using a demand forecast database (812); storing the transactional data pertaining to the virtual and current seasonal products and the stores using the transactional database (814); storing information pertaining to tracking of

product stock, suppliers, employees, purchase orders, sales or combination thereof using an inventory database (816); storing information pertaining to the climate or environmental conditions such as effect of the climate on the product, the sale of product in different seasons with varying climate or combination thereof using a store climate information database (818); and storing information pertaining to the promotion of the product such as personal selling, advertising, sales promotion, direct marketing, publicity strategy, using a promotion information database (820), wherein each of the database is communicatively coupled via a network (822).
13. The method as claimed in claim 12, further comprises of storing information pertaining to the future information of the quantity demanded in response to a one percent change in price of the product using the price elasticity database (810).
14. The method as claimed in claim 12, further comprises of storing information pertaining to quantity of a product or service that consumers will purchase in future using the demand forecast database (812).
15. The method as claimed in claim 12, further comprises of finding the impact related to the product such as climate effects, market response effects using the demand forecast database (812).
16. The method as claimed in claim 12, further comprises of storing information pertaining to profit information related to the product, the sales requirement in near future, orders of the product, all product activity records using the transactional database (814).
17. A system for demand forecasting of seasonal products for pricing during off-season period, characterized by demand forecasting in absence of the past transaction data and the presence of current year complete or partial in-season data, the system comprises of:

a. a virtual product and transaction data creation module (702)
adapted to create a plurality of virtual products and corresponding
transaction data based on current and historical product attributes
and current year in-season and historical transaction data;
b. a virtual product transaction and current year in-season data
normalization module (704) adapted to normalize the plurality of
virtual products, corresponding transaction and the current year in-
season transaction data and further re-scaling normalized current
year in-season transaction data as per peak-period and peak-sales
of virtual product;
c. a sales response function design and estimation module (706)
adapted to estimate the impact of at least one factor on the plurality
of virtual products and current season products at different
hierarchy level for representing non-linear behaviour of sales
response function for off-season period;
d. a sales response function updation module (708) adapted to
determining the weightage need to be considered for the plurality
of virtual product and corresponding transaction data of past years
for further deriving or updating the impact of each the factor on
off-season period; and
e. a demand forecasting module (710) adapted to forecast demand by
utilizing different hierarchies of the attributes, wherein the virtual
product and transaction data creation module (702); the virtual
product transaction and current year in-season data normalization
module (704); the sales response function design and estimation
module (706); the sales response function updation module (708);
and the demand forecasting module (710) are coupled to an
electronic hardware processor configured to perform operative
functions of each.
18. The system as claimed in claim 17, further comprises of a virtual bucket and clustering module (802) is adapted to cluster the similar virtual

products and current season products with similar factors and characteristics together inside a similar virtual bucket which may be further utilized to form a group of stores.
19. The system as claimed in claim 18, further comprises of a store group estimation module (804) is adapted to intake all the information of different stores from the virtual bucket database (806) and transactional database (814).
20. The system as claimed in claim 17, further comprises of a virtual bucket & clustering module (802) adapted to cluster similar products with similar factors based on their characteristics, together inside a similar virtual bucket and store the same in a virtual bucket database (806); a store group estimation module (804) adapted to take input of information pertaining to different stores from the virtual bucket database (806) and a transactional database (814); the sales response function design and estimation module (706) adapted to estimate the quantity of a product or service that consumers will purchase in future; a store group database (808) adapted to store information pertaining to the similar stores; a price elasticity database (810) adapted to store information pertaining to the store group estimation module (804), the sales response function design and estimation module (706) and the sales response function updation module (708); a demand forecast database (812) adapted to store information pertaining to the demand forecasting module (710); the transactional database (814) adapted to store the transactional data pertaining to the virtual and current seasonal products and the stores; an inventory database (816) adapted to store information pertaining to tracking of product stock, suppliers, employees, purchase orders, sales or combination thereof; a store climate information database (818) adapted to store information pertaining to the climate or environmental conditions such as effect of the climate on the product, the sale of product in different seasons with varying climate or combination thereof using; and a promotion information database (820)

adapted to store information pertaining to the promotion of the product such as personal selling, advertising, sales promotion, direct marketing, publicity strategy, wherein each of the database is communicatively coupled via a network (822).
21. The system as claimed in claim 20, wherein the price elasticity database (810) is further adapted to store information pertaining to the future information of the quantity demanded in response to a one percent change in price of the product.
22. The system as claimed in claim 20, wherein the demand forecast database (812) is further adapted to store information pertaining to quantity of a product or service that consumers will purchase in future and finding the impact related to the product such as climate effects, market response
effects.
23. The system as claimed in claim 20, wherein the transactional database
(814) is further adapted to store information pertaining to profit
information related to the product, the sales requirement in near future,
orders of the product, all product activity records.