DECISION TREE METHOD APPLIED IN ECONOMICS AND STATISTICS

Autor/autori: Rocsana BUCEA-MANEA-TONIS, Radu BUCEA-MANEA-TONIS

Rezumat: Arborii de decizie permit reprezentarea sub forma de diagrama a unor evenimente viitoare previzionate care conditioneaza decizia. Practic se reprezinta grafic toate combinatiile posibile ale variantelor decizionale si ale starilor sistemului in fiecare moment de timp. Deciziile sunt influentate de evenimente aleatoare a caror probabilitate poate fi anticipata. Reprezentarea vizuala a tuturor variantelor posibile faciliteaza luarea deciziilor, in timp real si foarte usor. In acest articol demonstram cum se implementeaza un arbore de decizie bazat pe algoritmul ID3. Deasemenea am implementat o simulare privind determinarea strategiei optime de distributie a unei companii pe baza unui arbore de decizie

Cuvinte cheie: entropie, castig de informatie, ID3, simulare

Abstract: Decision trees allow graphic representation of expected future events that condition the decision. In fact we plot all possible combinations of decision alternatives and states of the system each time. Decisions are influenced by random events whose probability can be predicted. The visual representation of all possible variants facilitate decision-making, in real time and very easily. In this paper we show how to implement a decision tree based on ID3 Algorithm. We also made a simulation in order to determine the optimal distribution strategy based on a decision tree

Keywords: entropy, information gain, ID3, simulation

DOWNLOAD PDF