Teburin Abubuwan Ciki
- 1. Gabatarwa
- 2. Babban Fahimta
- 3. Tsarin Tunani
- 4. Ƙarfi da Rauni
- 5. Bayanai Masu Amfani
- 6. Cikakkun Bayanai na Fasaha
- 7. Sakamakon Gwaji
- 8. Misalin Tsarin Bincike
- 9. Aikace-aikace na Gaba
- 10. Manazarta
1. Gabatarwa
Takardar "Rarraba Canjin Kuɗi ta Hanyar Ƙungiya" ta Fenn da sauransu (2010) ta yi amfani da kimiyyar sadarwa don nazarin dangantaka a kasuwar musayar kuɗi (FX) daga 1991 zuwa 2008. Marubutan sun gina hanyar sadarwar kasuwar FX inda nodes ke wakiltar farashin musayar kuɗi kuma edges masu nauyi ke wakiltar dangantaka masu dogaro da lokaci. Suna amfani da gano ƙungiyoyi masu canji don bayyana yadda farashin musayar kuɗi ke taruwa da yadda matsayinsu ke canjawa cikin lokaci. Wannan hanya tana ba da hangen nesa na tsaka-tsaki game da tsarin kasuwa, wanda ya wuce hanyoyin gargajiya na bishiyar mafi ƙarancin haɗi.
2. Babban Fahimta
Babban fahimta ita ce, farashin musayar kuɗi ba su da zaman kansu; suna samar da ƙungiyoyi masu canji waɗanda ke nuna matsayin aiki a kasuwar FX ta duniya. Misali, manyan kuɗi kamar USD, EUR, da JPY suna mamaye ƙungiyoyinsu akai-akai, yayin da kuɗin kasuwanni masu tasowa ke canza matsayi a lokutan rikici. Marubutan sun nuna cewa tsarin ƙungiya ba ya tsaya tsayin daka—yana canjawa tare da abubuwan kasuwa, kamar Rikicin Kuɗi na Asiya (1997) da gabatar da Euro (1999). Wannan yana ƙalubalantar zato na dangantaka mai tsayi a inganta fayil da sarrafa haɗari.
3. Tsarin Tunani
Takardar tana bin tsarin tunani bayyananne: (1) Gina hanyar sadarwa ta dangantaka na farashin musayar kuɗi 33 a cikin shekaru 18. (2) Aiwatar da gano ƙungiyoyi masu canji ta amfani da hanyar taga-lokaci. (3) Nazarin zama memba na ƙungiya da matsayin nodes. (4) Haɗa canje-canjen ƙungiya zuwa abubuwan kasuwa da aka sani. Marubutan suna amfani da nazarin node-centric don bin diddigin yadda farashin musayar kuɗi guda ɗaya ke motsawa tsakanin ƙungiyoyi, yana bayyana waɗanne kuɗi ne "shugabanni" (masu tsayi) da waɗanne ne "mabiya" (masu canji). Wannan tsarin yana da tsauri amma yana ɗauka cewa dangantaka daidai take da mu'amala—sauƙaƙawa wanda ke watsi da alaƙar dalili.
4. Ƙarfi da Rauni
Ƙarfi: Amfani da gano ƙungiyoyi masu canji ya kasance sabon abu a lokacin, yana ba da girman lokaci wanda hanyoyin MST masu tsayi ba su da shi. Bayanan (1991-2008) sun ƙunshi manyan abubuwan kasuwa, yana sa sakamakon ya kasance mai ƙarfi. Nazarin node-centric kayan aiki ne mai amfani don gano canje-canjen mamayar kuɗi.
Rauni: Ma'aunin dangantaka ya dogara da Pearson, wanda ke ɗauka daidaitaccen layi da al'ada—duka biyun sun keta a cikin dawowar FX. Algorithm ɗin gano ƙungiya (inganta modularity) yana da iyakokin ƙuduri da aka sani, yana iya rasa ƙananan ƙungiyoyi. Takardar kuma ba ta da gwaje-gwajen mahimmancin ƙididdiga don canje-canjen ƙungiya. Idan aka kwatanta da aikin baya kamar Barigozzi da sauransu (2011) akan hanyoyin sadarwa na dangantaka na ɓangare, wannan hanya ba ta da ƙwarewa wajen sarrafa dangantaka na karya.
5. Bayanai Masu Amfani
Ga masu aiki: (1) Yi amfani da zama memba na ƙungiya mai canji a matsayin alamar jagora don canje-canjen tsarin kuɗi. Misali, idan kuɗi kamar GBP ya motsa daga ƙungiyar da EUR ta mamaye zuwa ƙungiyar da USD ta mamaye, yana iya nuna canji a daidaitawar manufofin kuɗi. (2) Haɗa tsarin ƙungiya cikin rarraba fayil—saka hannun jari a kuɗi daga ƙungiyoyi daban-daban don rage haɗarin tsari. (3) Sa ido kan kwanciyar hankalin ƙungiya a matsayin alamar canji; ƙungiyoyi da suka wargaje sau da yawa suna gaban tashin hankalin kasuwa.
6. Cikakkun Bayanai na Fasaha
An gina hanyar sadarwa ta amfani da ma'aunin dangantaka na Pearson tsakanin dawowar log na farashin musayar kuɗi. Ga farashin musayar kuɗi biyu $i$ da $j$, dangantaka a lokaci $t$ ita ce:
$$\rho_{ij}(t) = \frac{\langle r_i r_j \rangle - \langle r_i \rangle \langle r_j \rangle}{\sigma_i \sigma_j}$$
inda $r_i$ shine dawowar log, $\langle \cdot \rangle$ yana nuna matsakaicin taga mai jujjuyawa, kuma $\sigma_i$ shine madaidaicin bambanci. An ƙididdige hanyar sadarwa don riƙe dangantaka masu mahimmanci kawai ($|\rho| > 0.3$). Gano ƙungiya yana amfani da algorithm na Louvain akan aikin modularity:
$$Q = \frac{1}{2m} \sum_{ij} \left[ A_{ij} - \frac{k_i k_j}{2m} \right] \delta(c_i, c_j)$$
inda $A_{ij}$ shine matrix ɗin haɗi, $k_i$ shine digiri, $m$ shine jimlar nauyin edge, kuma $\delta(c_i, c_j)$ shine 1 idan nodes suna cikin ƙungiya ɗaya.
7. Sakamakon Gwaji
Marubutan sun gano ƙungiyoyi 4-6 masu dawwama a cikin lokaci. Babban sakamako ya haɗa da:
- Hoto 3: Zama memba na ƙungiya a cikin lokaci yana nuna cewa USD, EUR, da JPY suna samar da tushe masu tsayi, yayin da AUD, NZD, da CAD ke canzawa tsakanin ƙungiyoyi.
- Hoto 5: Rikicin Kuɗi na Asiya (1997) ya haifar da wargajewar kuɗin Asiya na ɗan lokaci zuwa ƙungiya daban.
- Hoto 7: Gabatar da Euro a 1999 ya haifar da haɗewar kuɗin Turai zuwa ƙungiya ɗaya.
- Tebur 1: Nazarin node-centric ya nuna cewa USD yana da mafi girman "dawwamar ƙungiya" (0.85), yana nuna matsayinsa na mamaye.
8. Misalin Tsarin Bincike
Ka yi la'akari da manajan fayil wanda ke da hannun jari a EUR/USD, GBP/USD, da JPY/USD. Ta amfani da tsarin takardar, za ta:
- Ƙididdige dangantaka na watanni 6 masu jujjuyawa tsakanin waɗannan nau'ikan.
- Aiwatar da gano ƙungiya don ganin ko suna cikin ƙungiya ɗaya.
- Idan suna cikin ƙungiya ɗaya (misali, duk suna da alaƙa da Turai), za ta rage hannun jari don guje wa haɗarin tattarawa.
- Idan GBP ya koma ƙungiyar USD, za ta iya yin shinge daban.
Wannan misali ne mai sauƙi; ainihin bincike yana amfani da kuɗi 33 da tagogi masu canji.
9. Aikace-aikace na Gaba
Hanyar tana da faɗaɗa bayyananne: (1) Sa ido na ainihin lokaci: Tare da bayanan mita mai girma, ƙungiyoyi masu canji na iya zama tsarin gargaɗin farko don rikicin kuɗi. (2) Kasuwannin cryptocurrency: Aiwatar da tsarin guda ɗaya ga Bitcoin da altcoins na iya bayyana irin tsarin mamayar. (3) Hanyoyin sadarwa masu yawa: Haɗa hanyoyin sadarwa na dangantaka da hanyoyin sadarwa na kwararar kasuwanci ko riba na iya ba da fahimta mai zurfi. (4) Haɗin koyon inji: Amfani da graph neural networks don hasashen canje-canjen ƙungiya na iya inganta hasashen. Hanyar takardar tushe ce, ba samfurin ƙarshe ba.
10. Manazarta
- Fenn, D. J., Porter, M. A., Mucha, P. J., da sauransu. (2010). Rarraba Canjin Kuɗi ta Hanyar Ƙungiya. arXiv:0905.4912v2.
- Mantegna, R. N. (1999). Tsarin matsayi a kasuwannin kuɗi. European Physical Journal B, 11(1), 193-197.
- Barigozzi, M., Brownlees, C., & Engle, R. F. (2011). Sabuwar hanya don nazarin manyan hanyoyin sadarwa na kuɗi. Journal of Econometrics, 163(2), 188-204.
- Newman, M. E. J. (2006). Modularity da tsarin ƙungiya a hanyoyin sadarwa. Proceedings of the National Academy of Sciences, 103(23), 8577-8582.
- Blondel, V. D., Guillaume, J.-L., Lambiotte, R., & Lefebvre, E. (2008). Saurin buɗe ƙungiyoyi a manyan hanyoyin sadarwa. Journal of Statistical Mechanics, P10008.