Abstract:
"
In this research, the research and the implementation are on personalized content-based
packages recommendation for mobile telecommunication subscribers. In fact, the
subscriber will be recommended by items like the ones the subscriber preferred and
used in the past. For the recommendation, there is a higher probability of commitment
from the recent past data of 6 Months’ time. Further one year and 6 months data is also
committed towards the recommended results in a lesser probability rate. Accordingly,
the recommended results should be the same as the actual results. In this research, the
idea is to interpret a mobile content-based product recommender system for mobile
subscribers to enhance their facilities. The area covered with the recommender system
are recommender system in the telecom domain, considering data usage on different
platforms and various content-based packages, and Comparison of algorithms to come
up with the most matching result with the actual result.
Telecom operators provide their subscribers with useful services such as voice calls,
video calls, SMSs, and data facilities etc. without which we can hardly imagine our
lives anymore. Due to the huge product assortments and complex descriptions of
telecom products, it is a great challenge for customers to select appropriate products.
There are various service types and service items. With such a vast number of products
and so complex description, it is becoming increasingly difficult for customers to find
their favorite products quickly and accurately. It is important to develop a
recommendation approach/ initiative to support customers in the selection of the most
appropriate telecommunication products. The machine learning model built under this
research is an ideal solution for the mentioned issue which encountered by many people
in the world. To complete the research project, there are several tools and techniques
used and as supportive tools there are some other tools were used which have less
importance.
The prominent tools used to build main findings are MS Azure, Power BI, R Studio
and Python. As main findings of this research project, there are three"
