Data Science in Insurance

The world is creating huge volumes of data, with increasing data production exponentially in recent years. Every day, no less than 2.5 quadrillion bytes of data (i.e. 2500 million billion bytes) are produced, according to estimates in 2015. To demonstrate the rate of increase, it has been estimated that 90% of all the data existing at the time had been produced during the previous three years alone [1]. We are not only witnessing an increase in the amounts of data, but also the be collected in new ways:

* massive amounts of customer information collected via the Internet search engines such as Google;

* the increase in data created via social networking sites such as Facearticle and Twitter;

* data acquired by mobile phones, tablets and smart watches, and telecommunications- munications and accessories.

We are rapidly living in what is known as the Internet of Things, or IoT.

It is about the increase in the number of household appliances that are all connected to each other through networking and all create data and interact with us.

In the UK, the Internet of Things is gaining popularity at home heating systems, allowing more economical energy consumption in smart houses. With the increasing importance of wearable gadgets that monitor work-apart from routines and well-being statistics, technology is also revolutionizing fitness diets. Insurers have already seen potential due to these specificities examples applied to the home and insurance coverage. Thanks to technolog-ical progress and processing power, data can now be stored, manipulated, and analyzed much faster and at a lower cost than in the past.

Insurers work in an environment richer in data and encryption methods, where increased processing power allows computers to gather, convert and analyze data more efficiently. Data science and machine learning allow actuaries to develop established actuarial fields while adopting new approaches to improve various business activities, governance procedures and customers happiness. To be competitive in this rapidly evolving and demanding sector, insurers must invest in business intelligence.

1-Data scientists and actuaries

Insurance companies are increasingly demanding that their employees have data science skills. Machine learning often has three components: coding, computer studies and domain expertise. While IT allows the data translation and development of algorithms, mathematical principles allow it use data to build models and predict future events. In Addition, machine learning must be able to understand the actual events and regulations to tackling the real issues. As a result, data science covers the entire spectrum data management, not just machine learning and scientific techniques.

These qualities, combined with professional actuarial competence and regulatory-tory understanding, are becoming more and more desirable in the employment accountants.

Although computer science and parametric modeling have many simi- the larities, putting the actuarial professions in a good position to take advantage of them emerging approaches to data analysis, they differ in the way computer scientists and actuaries happen in reality. The main distinction comes when creating and by executing sensible solutions. Actuaries often use their domain expert-tise to choose relevant models before focusing on adjusting the characteristics that are appropriate to achieve the goal. On the other hand, data engineers spend time and effort testing several algorithms before calculating suitable algorithms model parameters. In addition, these fields differ because actuaries create economic models, but computer scientists frequently rely on external spescialists to grasp the components of knowledge. Therefore, there are deviations in how the hypotheses are validated, the characteristics are chosen and the adjustment of the model is evaluated.

2-Big Data: challenges and opportunities

The insurance industry is no stranger to discovering patterns in huge amounts of data. Nevertheless, this remains an important difficulty because the disorder-ganized and fluctuating data are difficult to analyze using typical technologies.

Big data requires creative technology and ways to capture, store and interloan massive volumes of data in addition to getting important business information.