Analytics for Basic Products in Mexico

Paul Millán, Félix Mata

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In this research, it was designed a methodology to forecast prices of products In the Mexican Republic. The dataset is composed of basic basket products. The work consists of analyzing open and mixed data of this dataset. The approach is centered on studying how is the behavior in time and location domains for three products, tuna, detergent, and milk. The data ranges for five years. Neural networks were used to analyze data, and several experiments of price forecast were issued using different granularity levels. The regression models were validated using two traditional approaches of the machine learning area, coefficient of determination, and mean absolute error. The experiments showed that the price of basic products varies by zone and it is possible to give a forecast with a percentage of 80% of precision.

Original languageEnglish
Title of host publicationTelematics and Computing - 8th International Congress, WITCOM 2019, Proceedings
EditorsMiguel Felix Mata-Rivera, Roberto Zagal-Flores, Cristian Barría-Huidobro
PublisherSpringer
Pages121-129
Number of pages9
ISBN (Print)9783030332280
DOIs
StatePublished - 2019
Event8th International Congress on Telematics and Computing, WITCOM 2019 - Merida, Mexico
Duration: 4 Nov 20198 Nov 2019

Publication series

NameCommunications in Computer and Information Science
Volume1053 CCIS
ISSN (Print)1865-0929
ISSN (Electronic)1865-0937

Conference

Conference8th International Congress on Telematics and Computing, WITCOM 2019
Country/TerritoryMexico
CityMerida
Period4/11/198/11/19

Keywords

  • Big Data
  • Data analytics
  • Data science
  • Forecasting
  • Machine learning

Fingerprint

Dive into the research topics of 'Analytics for Basic Products in Mexico'. Together they form a unique fingerprint.

Cite this