RAZISKAVA KREDITNEGA TVEGANJA S POMOČJO NEVRONSKIH MREŽ

Šenk, Bernarda

| | SLO | ENG | Cookies and privacy

Bigger font | Smaller font

First page > Show document

Show document

Title:	RAZISKAVA KREDITNEGA TVEGANJA S POMOČJO NEVRONSKIH MREŽ
Authors:	ID Šenk, Bernarda (Author) ID Mrkaić, Mićo (Mentor) More about this mentor... ID Avsec, Andreja (Comentor)
Files:	DR_Senk_Bernarda_2010.pdf (1,95 MB) MD5: CD861E32EEDD7335E5E00996140BC7C9 PID: 20.500.12556/dkum/f3d3893f-6421-44bb-8d7a-c50dcab70f4f
Language:	Slovenian
Work type:	Dissertation
Organization:	FOV - Faculty of Organizational Sciences in Kranj
Abstract:	Aktivno upravljanje s tveganji je nujno za vsako podjetje, katerega cilj je dolgoročni obstoj in konkurenčnost na trgu. Ignoriranje kreditnega tveganja oziroma premajhno posvečanje (pozornosti) temu problemu lahko podjetje pripelje v resne težave, še posebno v času recesije. Mehanizem odloga plačila omogoča dobavitelju nadzorovati finančno stabilnost kupca, kar je v primeru prodaj brez odloga plačila onemogočeno. Vsa podjetja se pri prodaji z odlogom plačila srečujejo z zamudami, ki so pri izpolnitvah pogodbenih obveznosti v poslovni praksi pogoste. V pričujočem delu analiziramo tiste informacije, ki podjetjem izboljšajo napovedi o verjetnosti in zamudi poravnavanju računov. Tega problema smo se lotili tako, da smo ocenjevali zamude z različnimi empiričnimi modeli. V raziskavo smo vključili kazalnike bonitete kupcev in osebnostne lastnosti njihovih odgovornih oseb ter predhodne zamude pri plačilih. Pokazalo se je, da so osebnostne lastnosti zelo pomembne v odnosu kupec — dobavitelj. V tem odnosu medsebojno zaupanje, ki temelji na vestnosti, lahko vodi do povečanega obsega sodelovanja. Raziskava je potrdila, da bodo osebe, ki so bolj vestne, z večjo verjetnostjo vestno izpolnjevale obveznosti. Če ima direktor višje izražen nivo vestnosti in s tem tudi faceto izpolnjevanje obveznosti, bo to podjetje svoje obveznosti plačevalo svojemu dobavitelju z večjo verjetnostjo v dogovorjenem roku plačila. Hkrati s tem pa si bo to podjetje bolj verjetno poskušalo izboljšati boniteto, posledično pa bo imelo lažji in cenejši dostop do finančnih virov pri bankah. Modelski rezultati kažejo, da sta najboljši napovedovalki zamud prvi dve predhodni zamudi, sledijo oblika podjetja, leto, produktivnost zaposlenega (dodana vrednost in skupni prihodki na zaposlenega), plačilni pogoji (dnevi vezave kratkoročnih poslovnih obveznosti in dnevi odloženega plačila), odprtost za izkušnje in izpolnjevanje obveznosti. Pri analizi kreditnega tveganja s pomočjo kazalnikov bonitete podjetij in faktorji bonitete je bila najboljša splošna regresijska nevronska mreža (GRNN). Za analizo povezave dimenzij osebnosti in poddimenzij vestnosti z zamudami pa je bila najboljša nevronska mreža z večplastnimi perceptroni (NN z MLP), ki vsebuje eno skrito plast. Ravno tako smo dobili najučinkovitejše ocenjevanje zamud, kjer so vključeni vsi faktorji in predhodne zamude, z linearno nevronsko mrežo, s splošno regresijsko nevronsko mrežo (GRNN) in nevronsko mrežo z večplastnimi perceptroni (NN z MLP). V vseh primerih pa je bila najslabša nevronska mreža z radialno osnovno funkcijo (RBFNN).
Keywords:	kreditno tveganje, nevronske mreže, boniteta, kazalniki, osebnost, vestnost, faktorska analiza, zamude.
Place of publishing:	Maribor
Year of publishing:	2008
PID:	20.500.12556/DKUM-15408
COBISS.SI-ID:	244176128
NUK URN:	URN:SI:UM:DK:LIVPK5SZ
Publication date in DKUM:	05.01.2011
Views:	3680
Downloads:	307
Metadata:
Categories:	FOV
:	ŠENK, Bernarda, 2008, RAZISKAVA KREDITNEGA TVEGANJA S POMOČJO NEVRONSKIH MREŽ [online]. Doctoral dissertation. Maribor. [Accessed 25 March 2025]. Retrieved from: https://dk.um.si/IzpisGradiva.php?lang=eng&id=15408 Copy citation

Average score:	0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 (0 votes)
Your score:	Voting is allowed only for logged in users.
Share:

Similar works from our repository:

Similar works from other repositories:

Hover the mouse pointer over a document title to show the abstract or click on the title to get all document metadata.

Secondary language

Language:	English
Title:	THE CREDIT RISK INVESTIGATION WITH NEURAL NETWORKS
Abstract:	Active risk management is necessary for every company whose goal is long-term existence and competitiveness on the market. Ignoring credit risk or devoting too little attention to this problem can get companies into serious trouble, especially in times of recession. The payment delay mechanism enables a supplier to control customer’s financial stability. If all sales were without delay, such control would be impossible. In business practice it is quite common that payments are late. In present work we tried to ascertain which pieces of information helps companies improve their forecasts of the probability of invoice settlement. We examine this problem using different empirical models. We used creditworthiness indexes, personality properties and previous delays as explanatory variables. Our results show that personality traits are play an important role in the customer-supplier relationship. In this relationship mutual confidence can lead to increased collaboration. Investigation has confirmed that firms employing more conscientious executives are more likely to fulfil their financial obligations. If the chief/financial executive officer has a higher level conscientiousness, his company will more likely pay its debt to its supplier in time. In addition, such companies are more likely to attempt to improve their creditworthiness and in order to reduce the cost of external financing. Empirical results show that the best predictors of payment delays are: the first and the second lag of prior delays, the legal form of the company, the year, employee productivity (value added per employee and total revenues per employee), payment condition (days payables outstanding – average days’ credit and payables deferral period), openness to experiences and dutifulness. General regression neural network (GRNN) was the best for credit risk analysis with creditworthiness indexes and creditworthiness factors. Neural network with multi layer perceptron (NN with MLP) containing one hidden layer was the best for linking analysis between the personality dimension and the conscientiousness subdimensions. We got the most efficient delay estimation including all factors and prior delays with the linear neural network, general regression neural network (GRNN) and neural network with multi layer perceptron (NN with MLP). Radial basis function neural network (RBFNN) was the worst in all cases.
Keywords:	credit risk, nevron networks, creditworthiness, indexes, personality, conscientiousness, factor analysis, delays.

Comments

Leave comment

You must log in to leave a comment.

Comments (0)

0 - 0 / 0

There are no comments!

Back