Neural Networks Within AI
After completing the reading assignment, submit a paper here that answer these questions:
1. For each of the significant problems you wrote about in Assignment A, questions 6/7, how useful are neural network and other ML/AI methods?
Transparency in the banking industry and more specifically in the foreign exchange market is the most significant problem. Exchange rates are beyond the domestic economy which puts the exporters and importers at the risk of unclear and adverse movements in the exchange rates. It is not easy to detect the times of trade as well as it liquidity. The analytics team should therefore have the ability to uncover data that is useful in detecting the market flow. Reinforcement machine learning will help the computer systems to detect all types of progress within the FX market. The Reinforcement learning algorithm gather feedbacks from users and build a graph containing all the new information arising related to the FX markets that are of interest to the bank. Through Artificial neural networks, non-linear mathematical models can be used to make predictions of behavior in high liquid markets hence ensuring efficiency and transparency in the FX markets.
2. In addition to the ones you wrote about in Assignment A, questions 6/7, are there any other business problems where you would want to use neural network and other ML/AI methods?
The financial system provides an environment where markets and trade bonds including stocks make use of high frequency data hence increasing the probability of risk. The banking industry faces the risk of fraud on a daily basis especially the foreign exchange sector which generates enormous data. Neural networks would be very essential in the detection of fraud through making both character and image recognition. The artificial neural networks take in multiple inputs and deduce hidden information and interwoven relationships. The ANN can help banks to detect whether different signatures have been written by one person. Replacing humans with artificial intelligence within the banking sector in solving large-scale optimization problems has minimized the economic burden associated with human error. Neural networks and machine learning are used to make analysis and predictions of FX trends in the market. With, AI it will be easy to develop liquidity searching algorithms and even make portfolio suggestions to clients.
3. For each of the instances mentioned in your answer to 1 and 2 above:
o What would be the variable/feature to be predicted and the variables that would help make the prediction? In enhancing transparency in the FX market, the variable to be predicted would be the amount made in the FX market by banks. The variables that would help in making the predictions would be times of trade and the liquidity of the market. In detecting fraud, the variables would be the probability of a committing a fraud, and the predictive feature would be change of signature and frequency rate to the bank.
o Where you would get the data you need and what challenges would you face in getting the data and doing the analytics (if any)? Many banks are beginning to appreciate the use of neural network and machine learning in aiding their every day performance and operations. ANN for example develops algorithms that aid in decision making. Big data storage systems. However, the biggest challenge with ANN is that it forms bias which may result in elimination of important segment of the data.
