Abstract:
Newspaper is a one of the key mass media news provider in Sri Lankan society.
Newspaper is perishable item that valued only one day. Therefore, newspaper
publishing companies need to print accurate quantity of copies of newspapers and
deliver the sales point on-time, maintain the minimum returns without oversupply or
shortage supply to the market. By this newspaper publishers can enjoy high
profitability and reduce unnecessary publishing cost. Therefore, companies need
smooth circulation process and accurate newspaper forecasting solution. This research
project main objective is to introduce a newspaper circulation system and sales
forecasting system in Sri Lankan newspaper industry. Research introduced a web base
solution for handing newspaper circulation that include newspaper issuing, returns
handling and mode of payment, Circulation system developed PHP scripting
language and MySQL used for database. Developing this Circulation system research
is used in Waterfall model. This research develop a newspaper sales forecasting
solution by using data mining techniques such as classification, prediction and time
series regression techniques which will be used to analyse and forecast newspaper
sales time dependent data points such as paper wise forecasting in next month issue,
customer wise issue, district wise issue, city wise issue etc; on the other hand,
newspaper company can identify movement of newspaper buyers between agencies,
seasonal changes in demand between newspapers and identify sales patterns regarding
to agencies, newspapers, district and nationally wise : This research use popular Weka
(Waikato Environment for Knowledge Analysis) software for time series analysis and
its allows to create forecasting models, evaluating and visualizing the results. Finally
this research developed Newspaper circulation system that handles by the main
circulation process. Evaluation of the sales forecasting this research identified the
Linear Regression is most suitable method for forecasting the newspaper sales
forecasting. However, Linear Regression and SMO Regression methods can be used
for forecast city wise and agent wise newspaper sales, as well as this research
identified the sales patterns that help to make management decision. By using this
system Newspaper companies may reduce their unnecessary cost relating to
Newspaper sales and distribution.