Assignment #4

Neural Networks are a non linear, adaptive systems that are inspired by the biological nervous system such as the brain. They can identify and learn correlating patterns between input data sets and corresponding target values. Neural networks can be trained to predict the outcome of new independent data. 

Credit card companies use neural networks to monitor card activity of the user. The network is able to develop a pattern of the users spending activity. From this it can figure out more about the person and that persons spending habits.This network develops a pattern and when that pattern seems broken from an unusual spending characteristic it is flagged. If a person normally spends between $5 and $200 on any given transaction and all of a sudden there is a $1500 charge out of state, the network will flag this as unusual activity. This benefits both the business and the consumer. The business such as Discover for example, is able to save money by preventing any further theft. The user is also benefited by saving money and possibly their credit score. This is a very important system that helps millions of users everyday. People fall victim to identity theft and credit card fraud everyday so this system has been implemented and is very precise in stopping fraud. 

Neural networks are new to weather prediction. Inputs would include weather reports from surrounding areas. Outputs would be the future weather in our area.  Even though we already have a current system of predicting the weather, the complexity of weather predictions using conventional computer analysis proves to be too demanding.  The use of a neural network, however, which learns rather than analyzing has shown to be very proficient in predicting local rainfall with upwards of 85% accuracy. This type of system can be a game changer for everyone involved.  Meteorologists are able to predict the weather with more accuracy and can be more confident in the predictions.  The weather in one area of the world can effect the weather in that of a different area. A pattern can be created from this. Weather is very unpredictable but a complex system such as neural networking can provide us with a more accurate insight on the future weather and make an impact on how to prepare for it.

http://udini.proquest.com/view/credit-card-fraud-detection-using-goid:304810406/

http://www.gizmodo.com.au/2011/12/the-neural-networks-watching-your-credit-card/

http://www.calsci.com/Weather.html