Today a number of products and solutions are driven by analytics. For instance, Amazon GO, recommendations systems, predictive keyboards used on smartphones and chatbot are few of the examples of solutions that are driven by analytics.
What is Business Analytics?
In God we trust;all others must bring Data.-Edwards Deming
Since, many years in the past, many corporations used to make decisions based mostly on opinions rather than decisions primarily based on correct data analysis. Opinion based decision making may be dangerous that often results in an incorrect decision.
The main goal of business analytics is to enhance the standard of decision making utilizing data analysis.
Statistical and operations analysis methods have been in use for several decades by many organizations however since the 21st century, organizations that use analytics have elevated exponentially. The main purpose of this increase in analytics is the theory of bounded rationality proposed by Herbert Simon(1972).
According to Herbert Simon, the growing complexity of enterprise issues, the existence of several various options, and the restricted time out there for decision making the demand for an extremely structured decision-making process utilizing past data for the effective management of organizations.
Business analytics is a set of statistical and operations analysis methods, artificial intelligence, information technology and management strategies used for framing a business problem,collecting data, and analysing the data to create value to organizations.
Information Technology is used for data capture, data storage, data preparation, data evaluation and data sharing. Today, a lot of the data are unstructured data – that’s not within the form of a matrix(rows and columns). Images, texts, voice, video, click stream are few of the examples of unstructured data. To analyse these data, you simply want to make use of Softwares such as R, Python, SAS, SPSS, Tableau, and so forth.
Data Science is likely one of the essential parts of analytics. It consists of statistical and operations analysis methods, machine learning and deep learning algorithm. The goal of data science is to identify the most appropriate statistical model/machine learning algorithms that can be used.
Business analytics is grouped into three varieties: descriptive analytics, predictive analytics and prescriptive analytics.
– Edward R. Tufte
“If the statistics are boring, then you have got the wrong numbers.”
Descriptive analytics is the simplest type of analytics that primarily makes use of simple descriptive statistics, data visualization techniques, and business-related queries to understand previous data.
The main goal of descriptive analytics is to give you innovative methods of data summarization. Descriptive analytics using visualization identifies trends within the data and you will be able to attach the dots to gain insights about associated businesses. Additionally, descriptive analytics makes use of descriptive statistics and queries to gain insights from the data.
– Ronald Coase
“If you torture the data long enough, it will confess”
Predictive analytics aims to predict the chance of occurrence of a future event such as forecasting demand for product and services, loan defaults, fraudulent transactions, insurance coverage claims and inventory market fluctuations.
While descriptive analytics is used for finding what has occurred previously, predictive analytics is used for predicting what's more likely to occur sooner or later.
“Every decision has a consequence”.– Damon Darrel
Prescriptive analytics is the highest stage of analytics that is used for selecting optimum actions as soon as an organization gains insights by descriptive and predictive analytics.
Prescriptive analytics will help you to find the optimum answer to an issue or in making the correct selection amongst a number of alternate options.
Techniques used in Descriptive, Predictive and Perspective Analytics
The methods used for descriptive analytics are Mean, median, mode, variance, Standard deviation, Skewness and Kurtosis.
Some of the commonly used methods used in predictive analytics are Regression, Logistic Regression, Classification Trees, Forecasting Techniques, K-nearest neighbours, Markov Chains, Random forest, boosting and neural networks.
The Tools used in Perspective analytics are linear programming, integer programming, multi-criteria decision-making models such as goal programming and analytical hierarchy process, combinatorial optimizations, non-linear programming and meta-heuristics.
Machine Learning Algorithms
Machine learning algorithms are part of Artificial Intelligence that imitates the human learning process, Human learns by a number of experiences to carry out a job. In the identical approach, a number of models are developed using machine learning algorithms and every model is equal to experience.
Mitchell (2006) outlined machine learning as follows:
"Machine learns with respect to a particular task T, perform metrics O following experience E, if the system reliably improves its performance P at task T, following experience E"
Assuming the task T to be a classification problem that is considering a buyer’s propensity to purchase a product. The performance P can be measured by a number of metrics such as general accuracy, sensitivity and area under the receiver operating characteristics curve(AUC). The experience E is analogous to different classifiers generated in machine learning algorithms such as Random forest – where a number of tees are generated and every tree is used for classification of a new case.
Machine Learning algorithms are categorised into four classes.
- Supervised Learning Algorithms: Supervised Learning Algorithms are used if the training dataset comprises both predictor (input) and outcome (output) variables. The learning is supervised by the fact that predictors(X) and the outcome (Y) are variables for the model to use. Regression, logistic regression, decision tree, random forest are among the methods used for supervised learning algorithms.
- Unsupervised learning Algorithms: Unsupervised Learning Algorithms are used when the training dataset has solely predictor (input) variables (X), however not the outcome variable (Y). Techniques such as K-means clustering and Hierarchical clustering are some of the examples of unsupervised learning algorithms.
- Reinforcement Learning Algorithms: If some instances, the input variable (X) and the output variable (Y) are unsure. One instance can be predictive keyboards or spell checks that you use in your smartphones. In these instances, Reinforcement Learning Algorithms are used. The algorithms are additionally utilized in sequential decision-making situations together with methods such as dynamic programming and Markov decision process.
- Evolutionary Learning Algorithms: These are algorithms that imitate the human learning process. They are used frequently to solve prescriptive analytics issues. The methods used are genetic algorithms and abt colony optimization.
To conclude, business analytics is an integral part of a corporation and nearly all of the decisions are made utilizing previous data. Innovation is an important thing to the success factor for analytics.