PREDIKCIJA CENE PROIZVODA NA OSNOVU OPISA NJEGOVIH KARAKTERISTIKA

  • Milana Bečejac
Ključne reči: Predikcija cene, Neuronske mreže, Karakteristike proizvoda, Word2Vector, BagOfWords, LSTM

Apstrakt

Tema ovog rada je rešavanje problema predikcije cene proizvoda na osnovu opisa njegovih karakteristika. Specificirani su, implementirani i verifikovani modeli, koji rešavaju ovaj problem. Namena ovih modela je pomoć kupcima u proceni da li je cena proizvoda adekvatna, kao i pomoć prodavcima u određivanju prikladne cene za svoj proizvod.

Reference

[1] Su, Beichen. An Application of Recurrent Neural Network: Prediction of Items’ Price by Description. Diss. UCLA, 2018.
[2] Kim, Kyoung-jae, and Hyunchul Ahn. "A recommender system using GA K-means clustering in an online shopping market." Expert systems with applications 34.2 (2008): 1200-1209.
[3] https://www.kaggle.com (pristupljeno u septembru 2019.)
[4] https://github.com/ChenglongChen/tensorflow-XNN/blob/master/doc/Mercari_Price_Suggesion_Competition_ChenglongChen_4th_Place.pdf (april 2019.)
[5] https://searchenterpriseai.techtarget.com/definition/recurrent-neural-networks (septembar 2019.)
[6] https://medium.com/mlrecipies/deep-learning-basics-gated-recurrent-unit-gru-1d8e9fae7280 (septembar 2019.)
[7] https://github.com/Sebastianvarv/MercariPriceSuggestion (maj 2019.)
[8] https://github.com/gdayal80/mercari/tree/local/Mercari%20Price%20Suggestion%20Challenge (maj 2019.)
[9] Raygada, Javier. "Product Price Suggestions for Online Marketplaces."
Objavljeno
2020-02-22
Sekcija
Elektrotehničko i računarsko inženjerstvo