1) Comparison of predictive, descriptive, and prescriptive analytics:
Predictive analytics is referred to as the process of using statistics and modeling techniques for explaining the future performance of the business or any program. Whereas, descriptive analytics is referred to as the process of examiningthe data or content that is performed in the past. Predictive and descriptive analytics are having oppositional objectives but they are closely related to each other (Selvaraj & Marudappa, 2018). The relation between these analytics is helped to get accurate information about the past and make better predictions for the future. The predictive analytic tools are used to fill the gaps in currently available data. It means Descriptive analytics is explaining “what happened in the past” and the predictive analysis is explaining “what might be happening in the future”. For example, Descriptive analytics is explaining “completed a study of student” and the predictive analytics is explaining “what job the student gets in future” (Jeble, Kumari, & Patil, 2016). The predictive analytics is taking historical data and using that data to highlighting the visible patterns of big data. The machine learning algorithm and various statistical models are used to gathering the correlations between the data. The best example for explaining predictive analytics is the student’s attendance list. Colleges are using the historical information of student data to predict whether the student gets 75% attendance or not (Mathur & Kaushik, 2014). So, this predictive analysis is mainly focus on the future of businesses.
Prescriptive data analytics is referred to as the process of using technology that helps to make better decisions with the help of raw data. Predictive analytics is providing the raw material that is used to make informed decisions in business. The predictive analytics is explaining “what will happen”, “When it will happen” and “why” and the prescriptive analytics are used to suggesting the options (KaterinaLepenioti, Bousdekis, Apostolou, & Mentzasa, 2020). The options are helped in reducing the future risks, outcomes for the taken decisions, and explain the advantages of future opportunities. Mainly the prescriptive analytics is limiting the processes that are for getting better future outcomes. Both of these predictive and prescriptive analytics are used to solve the problems that are coming from the predictive data analysis (Jeble, Kumari, & Patil, 2016). And the prescriptive analytics is providing “data-backed decision options” that are opposite to one another. These two are good options for making the decisions but the prescriptive analytics Maintainance is a more ideal solution or pain the trillion lines of data (Mathur & Kaushik, 2014). Both the prescriptive and prescriptive analytics are together to present a colorful visual.
Predictive, prescriptive and descriptive analytics are all necessary tools to be used in business strategy for the profit. While Predictive analytic looks at past data to determine likelihood of its future, descriptive model looks at history of the company to determine the present scenario outcomes, and prescriptive model utilizes technologies that has influence of artificial intelligence technique. It uses highly sophisticated process to understand the business strategy to determine future performance, based on current and historical data. We can also argue that where the cost of human error is high, perspective analytics comes into play (Halton, 2019).
Although new analytical tools such as prescriptive and predictive is gaining popular in disclosing solution for the business, there will always be a need for descriptive analytics as it provides solutions in an efficient format (Frankenfield, 2019).
Insurance and marketing utilize the help of predictive analytical tools as the manner of data involves studying past behavior to predict future, while descriptive analytic uses a range of data to have a holistic view of performance and trends on which to base business strategy. As prescriptive analytical utilizes artificial intelligence technique it is highly effective in making quick decision and reducing errors. For instance, automation booking and evaluation of cost effectiveness in a healthcare or logistic industry (Segal, 2019).
Artificial intelligence is one of the important technologies that are widely applied in various business areas to ensure that it evaluates the problem and provides better decision-making solutions to the managers. the process of artificial intelligence involves classification of data, splitting data into two sets one for training and the other for testing, training the data, developing prediction model and application of test to create accuracy and confusion matrix. all these activities will be performed automatically when artificial intelligence is applied. the data collected will be perfectly analysed and appropriate relationship will be established between different attributes. this ensures the dependency of target variable on independent variables that continuously generate data.
There are several differences between deep learning and machine learning. Usually when these two types of models are compared the factors that are considered to compare them are data requirement, accuracy come on time required for training the data, and resources required (Patidar, 2018). considering the data requirement, machine learning requires lesser data compared to deep learning to make predictions about the data. with the help of less data itself machine learning can make better assumptions and provide predictions that help a business manager to address their problems. The accuracy of deep learning is high as it uses larger data to make predictions about given problem. machine learning provides better accuracy but when compared to deep learning it is usually observed to be low. considering the size of the data that is required for performing analysis deep learning need larger time to train the data whereas machine learning need lesser time too both on the same task. from these comparisons we cannot conclude one is good or bad, based on the requirements and the kind of data we are analysing the decision about selection of these two technologies depends. both the tools have enough ability to provide appropriate analysis and develop prediction models.
Artificial Intelligence (AI) uses machine learning to imitate human intelligence. AI can be an interdisciplinary science with various approaches and a branch of computer science which is concerned with building smart machines capable of performing tasks that typically require human intelligence. The computer has to grasp how to respond to specific actions and, therefore, uses algorithms and historical data to create a propensity model (Artificial Intelligence, 2020).
AI works by mixing large amounts of data and intelligent algorithms, which allows the software to learn automatically from features in the data. AI is enabled and supported via various technologies which are
• Graphical Processing Units- These provide the heavy compute power which is required for iterative processing
• The Internet of Things- which generates massive data from devices
• API (Application Programming interfaces)- Portable Packages of code add AI functions to existing products and software (Artificial Intelligence, 2020).
Deep learning is just a subgroup of machine learning. Deep learning is technically similar to machine learning and functions (hence why the terms are sometimes loosely interchanged). However, its capabilities are different.
While basic machine learning models become steadily better at whatever their function is, they still need some direction. If an AI algorithm returns an defective prediction, an engineer must step in and accommodate. With a deep learning model, an algorithm can decide on its own if a forecast is accurate or not through its neural network (Grossfeld, 2020).