1. Gabatarwa
Kasuwar forex, wacce ke aiki awanni 24 a rana, tana nuna sauyi mai yawa da canje-canjen farashi masu mahimmanci. Masu ciniki galibi suna fuskantar kalubale saboda canje-canjen farashi marasa tsinkaya da hadarin asara. Wannan bincike yana ba da shawarar robot ciniki wanda ya dogara da hankali na fuzzy don inganta dabarun fasaha na yau da kullun, musamman ta amfani da alamomin RSI, CCI, da Stochastic, don inganta daidaiton ma'amala da riba.
2. Teburin Abubuwan Ciki
- 3. Hanyoyin Bincike
- 4. Sakamakon Gwaji
- 5. Bayanan Fasaha
- 6. Nazari: Samar da Siginar Ciniki
- 7. Hanyoyi na Gaba
- 8. Bincike na Asali
- 9. Manazarta
3. Hanyoyin Bincike
Tsarin da aka gabatar yana maye gurbin ma'aunin alama na gargajiya da tsarin tunani na fuzzy Mamdani ga kowace alama (RSI, CCI, Stochastic). Ana hada sakamakon ta hanyar tsarin zabe don samar da siginar saye/siyarwa.
3.1 Tsarin Tsarin Fuzzy Mamdani
Kowace alama ana fuzzification ta amfani da ayyukan membobin (misali, kadan, matsakaici, mai yawa). Tushen doka yana ayyana ayyukan ciniki bisa ga hadewar wadannan saitin fuzzy. Misali, idan RSI ya yi kadan kuma CCI ya yi kadan, to siginar ita ce saye mai karfi.
3.2 Fuzzification na Alamomi
RSI an tsara shi zuwa saitin fuzzy: an sayar da yawa (0-30), tsaka-tsaki (30-70), an saya da yawa (70-100). CCI yana amfani da jeri iri daya, yayin da Stochastic %K da %D aka fuzzification zuwa yankuna kadan, matsakaici, da mai yawa.
3.3 Tsarin Zabe
Kowane tsarin fuzzy yana fitar da sigina (saye, siyarwa, rike). Kuri'a mafi yawa tana tantance aikin karshe. Wannan yana rage siginonin karya kuma yana inganta karfi.
4. Sakamakon Gwaji
An gwada robot na fuzzy akan bayanan forex na tarihi (EUR/USD, GBP/USD) na tsawon watanni 12. An kwatanta aikin da hanyoyin gargajiya guda uku: (1) RSI na yau da kullun, (2) CCI na yau da kullun, (3) Stochastic na yau da kullun.
4.1 Kwatanta Ma'aunin Riba
Robot na fuzzy ya sami ma'aunin riba na 2.45, idan aka kwatanta da 1.78 (RSI), 1.65 (CCI), da 1.92 (Stochastic). Wannan yana wakiltar inganta 27% akan mafi kyawun hanyar gargajiya.
4.2 Binciken Riba da Asara
Ribar net ta karu da 35% idan aka kwatanta da matsakaicin hanyoyin gargajiya. An rage yawan asara zuwa 12%, idan aka kwatanta da 18% na RSI na gargajiya da 22% na CCI na gargajiya.
5. Bayanan Fasaha
Tsarin tunani na fuzzy yana amfani da tsarin lissafi mai zuwa don karfin sigina $S$:
$S = \frac{\sum_{i=1}^{n} w_i \cdot \mu_i}{\sum_{i=1}^{n} w_i}$
inda $w_i$ su ne nauyin doka kuma $\mu_i$ su ne matakan membobin. Defuzzification yana amfani da hanyar tsakiya.
6. Nazari: Samar da Siginar Ciniki
Yi la'akari da yanayin da RSI = 25 (an sayar da yawa), CCI = -120 (an sayar da yawa), Stochastic %K = 15 (kadan). Tsarin fuzzy yana fitarwa: RSI fuzzy = saye mai karfi, CCI fuzzy = saye mai karfi, Stochastic fuzzy = saye. Zabe yana haifar da siginar saye mai karfi. Wannan yana guje wa siginonin rikici na hanyar gargajiya.
7. Hanyoyi na Gaba
Aiki na gaba zai iya hada tsarin fuzzy masu koyo wadanda ke koya daga sauyin kasuwa, hada nazarin jin daga labarai, ko amfani da koyon karfafawa mai zurfi don inganta nauyin doka. Hakanan za'a iya fadada tsarin zuwa kasuwannin cryptocurrency.
8. Bincike na Asali
Mahimmanci: Wannan takarda ta nuna cewa hankali na fuzzy zai iya inganta alamomin fasaha na gargajiya sosai ta hanyar magance rashin tabbas na kasuwa, wanda ya kai ga inganta ma'aunin riba da 27%. Tsarin zabe yana rage siginonin karya, wani lahani na yau da kullun a cikin tsarin da ya dogara da ma'auni mai tsauri.
Tsarin Tunani: Marubutan sun fara da matsalar sauyin kasuwa, sun ba da shawarar tsarin fuzzy Mamdani ga kowace alama, sun hada sakamakon ta hanyar zabe, kuma sun tabbatar da hanyoyin gargajiya. Tsarin yana da ma'ana amma ya rasa gwajin kididdiga mai tsauri (misali, ma'aunin Sharpe, simulation na Monte Carlo).
Karfi da Rashin Karfi: Karfi sun hada da hanyar bincike bayyananniya da inganta aiki. Rashin karfi: binciken yana amfani da alamomi uku kawai da nau'in kudi guda daya; yiwuwar overfitting yana nan. Takarda ba ta tattauna farashin ma'amala ko slippage ba, wanda zai iya rage riba. Kamar yadda Goodfellow et al. (2016) suka lura a cikin Deep Learning, dogaro da yawa akan tsarin tarihi na iya haifar da rashin iya gaba a kasuwanni marasa tsayayye.
Abubuwan Aiki: Masu ciniki ya kamata su yi la'akari da tsarin hadewa na fuzzy-gargajiya don siginonin shiga. Duk da haka, dole ne su gwada a kan yanayin kasuwa da yawa (masu tafiya, masu kewayo, masu sauyi) kuma su hada da sarrafa hadari (misali, asarar tasha, girman matsayi). Hanyar fuzzy tana da alƙawari amma ba maganin dukkan matsaloli ba.
9. Manazarta
- M. Shabani, A. Nasiri, da H. Nafardi, "Forex Trading Robot Using Fuzzy Logic," 2023.
- J. J. Murphy, Technical Analysis of the Financial Markets, New York Institute of Finance, 1999.
- L. A. Zadeh, "Fuzzy sets," Information and Control, vol. 8, no. 3, pp. 338–353, 1965.
- I. Goodfellow, Y. Bengio, da A. Courville, Deep Learning, MIT Press, 2016.
- J. S. R. Jang, "ANFIS: adaptive-network-based fuzzy inference system," IEEE Trans. Syst., Man, Cybern., vol. 23, no. 3, pp. 665–685, 1993.