Você está na página 1de 30

Applications of Intelligent

Technologies in
Retail Marketing
Azka Akram
BE-16-59
 Retailing:
Sale of goods and services from businesse
s to an end user.
 Retail Marketing:

The process by which retailers promote a


wareness and interest of their goods and services in a
n effort to generate sales from their customers.
 Intelligent technologies for database analys
is have significantly impacted on design and
development of Decision Support Systems a
nd Expert Systems in diverse disciplines.
 A Decision Support System is an interacti
ve system that helps decision makers utilize
data and models to solve unstructured and s
emi-structured problems.
 An Expert System is a problem solving co
mputer program that achieves good perform
ance in a specialized problem domain that is
considered difficult and requires specialized
knowledge and skills.
 Both seek to improve the quality of decision
.
 Objective of the presentation
 Retail Technology an ever changing beast.
 Intelligent technology and retailers.
 Imagine a store with eyes and ears that feeds al
l important data back to you daily.
 Imagine knowing what the shopper had bought
and what he is looking for today.
 Valley Example
 Valley is a marketing and advertisement fir
m that is working with different brands and
retailers
 The brands they are working with are Adida
s, Audi, Levi’s, silver cross and Sony etc
 Valley and Audi’s A1 model showcasing th
roughout UK
Five Classes of Intelligent
Technologies
1. Fuzzy logic
2. Neutral networks
3. Soft computing
4. Case-based reasoning
5. Collaborative filtering
Fuzzy logic
 Basic idea of fuzzy logic and definition
 Development of the ability of human mind r
easoning and decision making into our mac
hinery through fuzzy logic
 Advantages and Disadvantages
Retailing Applications
1.Fuzzy Market Segmentation :
Clustering of customers into groups
2.Fuzzy Control-based Quick Response(QR)Reorder Sc
hemes for Seasonal Apparel:
Hung et al. employed a fuzzy controller to develop a no
vel and intelligent Quick Response reorder scheme
3.Fuzzy Intelligent Agents for Targeted E-Tailing:
Yager has demonstrated that fuzzy controller-based int
elligent agents can autonomously and instantaneously pers
onalize the display of advertisements on the basis of the vie
wer’s characteristics
4. Fuzzy Logic-based Prediction of Custom
er Churn:
Casabayo et al. employed a fuzzy logic ba
sed classification model to predict customer s
witch if a new retailer opens an outlet and the
prediction was 90% accurate in identifying the
customers who would switch
Neural Network and soft comp
uting

Tasmania Hina
Roll no. 39
Neural Network
 Definition;

“Neural networks are set of ‘Algorithms’ mo


deled loosely after the human brain, that are
designed to recognize patterns”.
Retailing Applications
(NN)
1.

2. Retail Marketing Segmentation.

3. Retail Forecasting Using Neural Networks.

4. Modeling Repeat purchase in Mail-order Ret


ailing.
Soft computing
 Definition;
“Soft computing is a collection of
techniques which help to construct a compu
tationally intelligent system”.
Proposed by L.A. Zadeh
( Inventor of Fuzzy logic).
Computationally Intelligent Sys
tem

 It is something which must posses Human li


ke expertise.
 They must be able to adapt and learn.
 Should be able to make decision and take ac
tion.
Characteristics of Soft computi
ng

 Does not required any mathematical model.


 No precise solution.
 Adaptive Algorithms (Adjust in Dynamic E
nvironment) as you can have different para
meters and inputs.
Retailing Applications
(Soft computing)
 Fuzzy Neural Network for modeling Apparel
Operations.

 What will happen if the retail season invent


ory is reduce?

 What will happen if we use a poor forecast


of stock keeping unit mix and/or demand vol
ume?
Continued

 What will be the impact if selling season


s are made shorter?

 What cost/ benefits will result if reorder l


ead times are reduce?
Continued

 Soft Computing-based Multi-agent Retailin


g Decision Support System.
 Bayesian Neural Network for Repeat Purch
ase Modeling in Mail-order Marketing.
 Soft Computing for Customer Targeting in
Database Marketing.
Tuba Ahmed
Roll # 26
Case Based Reasoning(CBR)

 Reasoning by remembering
 Solves new problem by adapting solutio
ns that were used to solve old problems
 Approach to problem solving and learni
ng
 Uses “k-nearest neighbor” technique
Retailing Application of CBR

CBR-Based Promotion Response Forecasting


It involves 3 steps
 Selects promotions from historical data

 Adjusts the sales of each analogous promotion

 Combines the forecast given by multiple cases


Pros & Cons of CBR
 Advantages
 Good when data appears as cases and when dat
aset is small
 Solutions are quickly proposed
 Disadvantages
 Cannot be applied to large datasets
 Poor in generalization
Collaborative Filtering

 Collaborative filtering is method of mak


ing automatic predictions about the inter
est of a user by collecting taste informati
on from many users
Retailing Application of CF

 CF-Based Market Basket Analysis


 Collaborative filtering is an approach that i
s used to perform market basket analysis
 In market basket analysis, we analyze:

“purchases that commonly happen togeth


er”
Pros & Cons of CF
 Advantages
 can be used to generate personalized reco

mmendations for users


 Disadvantages

 User profiles are required; when new prod


ucts are launched no ratings from users are
available
Future Research Directions

There are some research streams:


 Costumer Churn Prediction

 Repeat Purchase Modeling

 Construction of “ensemble” classifiers

 Exploit the emerging paradigm of evolving


connectionist systems
Conclusion:
This chapter focus on
 Several kinds of decision making problems

that commonly arise in retail marketing


 The use of RFID technology in retail chain

s
 Retail applications awaiting creative adapti
on and implementation of the exciting new
technologies

Você também pode gostar