SPECIALIST
Total years of experience :6 years, 7 Months
Worked on front-end development to design the pages of websites
Make the databases with ASP .NET MVC4 Framework
To ensure the database and handling all the backend queries to retrieve the data
To work in visual studio with C# for developing desktop applications
supply chain management and the manufacturing of the products. So we need to know the accurate
forecasting of retail sales. It is a difficult task to develop an accurate forecasting model because sometimes
we face under-forecasting and over-forecasting. Deep learning algorithms are very effective and useful for
making future predictions. These algorithms use the historical data as an input and provide the output in
the form of predictions. We used four different deep learning algorithms named CNN, MLP, LSTM, CNN-
LSTM with additional four dense layers for this purpose and choose the best one on the basis of
performance matrices. We gathered the one-year sales data of the pharmaceutical industry for training and
testing of our algorithms. We analyzed these four algorithms on the collected dataset and concluded that
the CNN-LSTM is the best fit for our dataset for the prediction of sales. CNN-LSTM gives the 0.06514 RMSE
and 0.29392 MSE error rates. These are the minimum error rates among the other three algorithms.
Supervisor: Asst. Prof. Mis Hina Kirn
Worked as front-end and backend
Thesis: Comparative Analysis of Deep Learning Algorithms for Retail Sale Prediction Today’s sales forecasting is the hot topic and demand of every business. It shows the importance between
URL removed due to policy violation. Please contact support for further information.