Entropy in Real-World Datasets and Its Impact on Machine Learning

• Other Authors: Kozak, Jan (Editor), Juszczuk, Przemysław (Editor)
• Format: Electronic Book Chapter
• Language: English
• Published: Basel MDPI - Multidisciplinary Digital Publishing Institute 2023
• Subjects: Information technology industries; Computer science; machine learning; optical networks; imbalanced data; one-class classification; entropy measure; real-world data; preprocessing; decision table; classification; query set; decision tree; differential cryptanalysis; metaheuristics; symmetric block ciphers; memetic algorithms; DES; simulated annealing; COVID-19; vaccination; agent-based modelling; dynamic stochastic general equilibrium models; scenario analyses; validation of results; stock index forecasting; CEEMDAN; ADF; ARMA; LSTM; hybrid model; classification measure; quality of classification; quality measure; preference-driven classification; fast iterative filtering; parameter adaptive refined composite multiscale fluctuation-based dispersion entropy; rotating machinery; fault diagnosis; short-term demand prediction; bike-sharing; travel characteristics analysis; hybrid TCN-GRU model; distributed data; decision tables; information systems; decision trees; decision rules; tests; reducts; association rules; Pawlak conflict analysis model; independent data sources; coalitions; dispersed data; length; support; greedy heuristics; feature selection; rough sets; action units; automatic translation; sign language; entropy of real data
• Online Access: DOAB: download the publication; DOAB: description of the publication

The topic of the reprint is very important nowadays, because ever-evolving machine learning techniques make it possible to obtain better real-world data. Therefore, this reprint contains information related to real data in fields such as automatic sign language translation, bike-sharing travel characteristics, stock index, sports data, fake news data, and more. However, it should be noted that the reprint also contains a lot of information on new developments in machine learning, new algorithms, algorithm modifications, and a new measure of classification quality assessment that also takes into account the preferences of the decision maker.