applications in Tertiary Sector

Machine Learning applications in Tertiary Sector-part 2

Posted on Posted in Machine Learning

This blog is a continuation of the last blog post regarding machine learning applications in tertiary sector of the economy.

Detailed below are some more machine learning applications in Tertiary Sector.

Hotel and Tourism

Artificial neural networks, Gaussian process regression and support vector machines are the most widely used machine learning methods for forecasting in this sector. These techniques help in suggesting mid and long-term forecasts by evaluating demand with seasonal fluctuations. These approaches are also used in hotel site evaluation by designing automated web Geographical Information System (GIS) applications. Hotel location selector and analysis tool set predicts various business success indicators associated with a location for example in Beijing, the tool calculates and visualizes desirable sites contingent on specific features of the hotel . For shorter forecast periods, methods help in improving standard with a direct approach and introduction of tools to automate information discovery process. The issue with these approaches is exploratory data analysis need to extract information from seasonal, cyclical and trend components. It is time-consuming and requires domain experts.

Banks

In the banking industry, self-improving algorithms detect fraud, content and information extraction, establishing identity, efficient credit card transactions, correct decision-making and even to predict credit scores for ascertaining bad loans. According to IBM research, in the US, the cost of fraud is $80 billion dollars annually with $2.4 billion lost through credit cards. Management of credit cards efficiently and effectively helps to uncover fraudulent transactions and immediately trigger a card rejection at a point of sales. For example, Feedzai utilizes machine learning models to detect fraud up to 30% earlier than traditional methods. Kabbage applies predictive analysis to offer investment to small businesses and consumers through automated lending platform. LendUp uses algorithms to pinpoint customers who are most likely to repay their loans. Allowing data processing enhances decision-making for companies such as Zest Finance. Billguard applies models and techniques to alert users to bad charge. DataMinr helps in real-time information discovery and Alphasense employs algorithms for high-frequency trading by identifying true signals automatically among massive amounts of data. In 2014,more than 40% of new hedge funds used computer models for their trades. Moreover, consulting firm A.T. Kearney reckons that robot-advisors will handle up to $2.2 trillion by 2020 .

Other applications in Tertiary Sector include:

Healthcare

In recent years, there has been an unprecedented rise in availability and size of clinical data. This presents an opportunity to develop various approaches better-personalised care. Similarly, dose-response models and disease progression models characterized by states that evolve over time help to treat individual patient best possible way. Pharmaceutical companies use data examination to deepen their understanding of diseases and design better products for treatments. Machine learning also helps regulators and policy makers to define policies that increase healthcare value and safety. For example, Enlitic uses deep learning and image analysis to make diagnostics faster, more accurate and more accessible. Ginger.io uses a combination of smartphone technology, data science and clinical services to create a personalized affordable way to deliver mental health care. Medaware employs big data analytics and algorithms to cut healthcare cost by identifying and preventing prescription errors in real-time. And Lumiata applies real-time analytics to help hospital networks and insurance carriers in providing higher quality care to more patients in less time.

Law

Machine learning acts as a compliment to legal counselling and helps lawyers save time and money by suggesting future legal outcomes against data from past client scenarios and other relevant public and private data. For instance, Supreme Court Ideal Point Miner is 76.9 % accurate for making predictions about US Supreme Court judge’s individual judicial preferences and voting behaviour.  It also aids close the justice gap by offering digital advice to those who cannot afford a lawyer. It offers dispute resolution platforms to help mediate between users. It also cuts the amount of time invested in legal research. An example is a use of language analysis by Ravel to check a case, pinpoint a particular case it refers to and figure out what ties the cases together.

Leave a Reply

Your email address will not be published. Required fields are marked *