Abstract
This chapter provides an overview of some popular business and industrial applications of machine learning/data mining algorithms. The survey-like introductory section provides a brief overview of some relevant historical and trending applications, while the other four sections present specific details on four selected business and industrial applications. Each section focuses on a different considered algorithm, namely neural networks, rule induction, tree algorithms, and neighborhood-based algorithms.
TopIntroduction
Companies around the world are constantly searching for ways to increase their returns and reduce costs. Many business problems can be formalized and described in the language of mathematicians or computer scientists, i.e., by using concepts such as time-series, graphs, or matrices. Machine learning (ML) can be used in many business-related cases. For example, a retailing company can solve many issues by means of ML and optimization: assortment selection, physical organization of products, making suggested offers to known customers based on their purchase history, inventory organization, segmenting buyers, predicting shopping missions, etc. Similarly, industrial companies might want to reduce their costs and increase their revenue. By using historical observations, and employing ML regression/classification, they can predict future expenditures and sales, which can enhance their decision-making potential. Moreover, robotic systems, which are now becoming standard for many industrial companies world-wide, use state-of-the-art ML methods, which further improves their efficiency and consequently produces more revenue with lower costs.
The presented overview is non-exhaustive and represents only a sample of all relevant applications of ML in business and industry, organized by categories. The following text will first give a more detailed insight of recent ML applications in the four selected business/industrial contexts:
After this, four specific business and industrial applications will be discussed in much more detail. More precisely, it will be presented how each of the following four selected ML algorithms is applied:
- 1.
neural network algorithm for wind speed prediction,
- 2.
tree-like algorithm for life insurance risk prediction,
- 3.
rule-induction algorithm for business intelligence,
- 4.
neighborhood algorithm for evaluating microstructure in metal alloy.
Key Terms in this Chapter
RNN: Recurrent neural networks suitable for NLP.
Cross Validation: Machine learning technique for evaluating performance of a model using labeled data.
NLP: Natural language processing.
PCA: Principal component analysis that transforms data in sub-dimensional space.