K-nearest neigbour method was used to identify factors that influence customer’s acceptance of personal loan, and therefore plan a follow-up targeted campaign to increase the number of borrowers. The data included 5,000 customers and 12 parameters. The method was implemented in R.
A sentiment analysis of customer’s review was performed on baby products that were sold in amazon. More than 183,000 reviews were classified as positive or negative based on the content of the review. The analysis was performed using GraphLab
The aim of the project was to predict handwritten numbers using neural network technique. A 5,000 by 401 pixel of handwritten data was provided. I used a neural network structure with three layers (with layer size of 400, 25, 10 ). The cost function and gradient of neural network was written and the weight was predicted using fmincg Octave algorithm.
Association rule was used to determine the precedence of online courses based on customers experience. The analysis was performed in R
Multiple linear regression model was used to identify factors that influence slump flow of high performance concrete. R was used to implement the model.
Multiple linear regression model was used to predict house price. GraphLib was used to implement the model.
Markov chain was used to develop a model that forecasts the short term and long term market share of an insurance company. The current market share and transition probabilities of a one-step transition matrix (that can be obtained from customer survey or similar market search) were input parameters. This model is a general purpose model and can be adapted for any other businesses given similar input parameters.
Stochastic modeling was used to compare the economic and environmental impact of energy consumption behaviour on household's sustainable energy systems, such as solar panel. I considered four scenarios: households that consume energy (1) during the daytime, (2) in the evening and early morning, (3) with high variation and (4) with low variation. The data for each scenario was simulated from a Gamma probability distribution assuming that all groups have the same average energy consumption.
A simulation study was carried out with a model developed in Arena, a flow oriented simulation software program. The optimum capacity for all key resources (doctors, nurses, ambulances etc) of the emergency healthcare system was identified and differnet options were tested and recommended to deal with crisis situations.
The objective is to analyze the current situation and improve the existing facility by utilizing low cost changes to the facility layout for proper space optimization and reducing traffic congestion. An implementation plan with work break down structure and schedule, execution plan for monitoring and control, risk assessment and financial plan with budget were designed for the project.
Developed a self-service doctor appointment scheduling application using Java.