Abstract:
Recommender System are
data is growing rapidly therefore these Information solutions could be capable
other recommendations to facilitate the
because our
works on the existing data pattern and predict some
still looking for personalized user decisions. Grocery recommendation system
recommendations for the customers which will enhance the customer personalized shopping
experiences while shopping from a Grocery Mart. The major issue exists in all recommender
systems is the Cold start problem. The cold start problem occurs when no data found from
oftwo types either Visitor cold past data to process the request. The cold start problem
customer added into the system as well as when a new product
are
start problem when a new
added into the system it is called as Product cold start problem in both case the recommender
systems are unable to generate the recommendations.
This thesis gives the implementation of hybrid filtering recommendations approach
to generate the recommendations based on personalized grocery shopping experiences during
shopping in a Grocery Mart. It also implemented the solution of cold start problem for both
either visitor cold start or product cold start by fetching the IoT based promotion cases
recommendations for those products. To generate the IoT based recommendations, we have
implemented a mobile application, Grocery Picker, which is interacting with the smart IoT
based advertisement solution using ESP8266 and display it onto the customer mobile.
This Grocery Picker application uses machine learning components that generate the
Recommendations based on the needs of customers. The main contribution ofthis thesis is
to provide the solution of cold start problem by the implementations of IoT based
|
recommendations using the smart advertisement component.