Teburin Abubuwan Ciki
- 1. Gabatarwa
- 2. Babban Fahimta: Binciken Masana
- 3. Tsarin Ma'ana na Takardar
- 4. Ƙarfi da Rashi
- 5. Bayanai Masu Aiki
- 6. Cikakkun Bayanai na Fasaha da Tsarin Lissafi
- 7. Sakamakon Gwaji da Bayanin Zane
- 8. Tsarin Bincike: Nazarin Shari'a
- 9. Aikace-aikace na Gaba da Ra'ayi
- 10. Bincike na Asali (Kalmomi 300-600)
- 11. Manazarta
1. Gabatarwa
Wannan takarda ta Pinčák da Bartoš (2015) ta gabatar da wata sabuwar hanya ga hasashen kasuwar forex ta hanyar yin samfurin motsin farashin musaya a matsayin tsinkaya akan wani tsari mai kama da kirtani. Marubutan suna jayayya cewa samfuran tattalin arziki na gargajiya sun gaza saboda yin watsi da farashin ciniki, damar arbitrage, da kuma yanayin rashin daidaito na bayanan kuɗi na gaske. Suna ba da shawarar tsarin kirtani da yawa don gina samfuran hasashen da suka dace da dabarun zaɓin fayil masu ƙarfi.
2. Babban Fahimta: Binciken Masana
Babban Fahimta: Takardar tana ƙalubalantar tushen zato na tattalin arziki na yau da kullun—wato, cewa za a iya ɗaukar jerin lokutan kuɗi a matsayin ci gaba, matakai na Gaussian. A maimakon haka, tana ɗauka cewa kasuwar forex tana da tsarin geometric na asali wanda aka fi kwatanta da ra'ayoyin ka'idar kirtani. Wannan babban canji ne daga motsin Brownian da samfuran tsarin Lévy.
3. Tsarin Ma'ana na Takardar
Marubutan sun fara da sukar samfuran da ake da su (Sashe na I), sannan suka gabatar da tsarin kirtani a matsayin sabon wakilcin motsin farashin musaya. Suna nuna yadda wannan tsarin ke ba da damar samfuran hasashen da suka dace kuma suna kwatanta hanyarsu da dabarun bin yanayi. Takardar ta ƙare da tabbatarwa ta zahiri akan bayanan kasuwar OANDA, wanda ya tabbatar da kwanciyar hankali na algorithm a tsawon lokutan ciniki masu tsawo.
4. Ƙarfi da Rashi
Ƙarfi: Takardar ta magance wani gibi na gaske—gazawar samfuran gargajiya don yin la'akari da farashin ciniki da bayanan marasa daidaito. Amfani da ka'idar kirtani sabon abu ne kuma yana ba da sabon hangen nesa kan tsarin kasuwa.
Rashi: Takardar ba ta da cikakken tsarin lissafi na samfurin kirtani. Tabbatarwa ta zahiri ta iyakance ga kasuwa ɗaya (OANDA) da lokaci ɗaya. Ba a kwatanta da samfuran koyon injin na zamani (misali, LSTM, Transformer) ba. Manufar 'tsarin kirtani' ta kasance ba a bayyana ta sosai ba.
5. Bayanai Masu Aiki
Ga masu aiki: Yi la'akari da haɗa bayanan marasa daidaito da farashin ciniki a cikin samfuran ku. Ga masu bincike: Bincika mahadar nazarin bayanan topological da jerin lokutan kuɗi. Samfurin kirtani na iya zama mafari ga hanyoyin geometric mafi rikitarwa kamar homology mai dorewa.
6. Cikakkun Bayanai na Fasaha da Tsarin Lissafi
Marubutan suna yin samfurin motsin farashin musaya a matsayin tsinkaya akan wani nau'in kirtani. Ma'auni mai mahimmanci shine:
$X(t) = \sum_{i=1}^{N} \alpha_i \phi_i(t) + \epsilon(t)$
inda $\phi_i(t)$ ayyuka ne na tushe da aka samu daga ka'idar kirtani, $\alpha_i$ sune coefficients, kuma $\epsilon(t)$ hayaniya ce. An ayyana tsarin kirtani da yawa kamar:
$S = \{s_1, s_2, ..., s_M\}$ inda kowane $s_j$ kirtani ce mai wakiltar wani tsarin kasuwa daban.
7. Sakamakon Gwaji da Bayanin Zane
Zane na 1: Wani zane na tsarin kirtani yana nuna hanyoyin farashin musaya a matsayin kirtani a cikin tsinkaya mai girma biyu. Zanen yana nuna yadda kirtani da yawa (masu wakiltar tsarin kasuwa daban-daban) ke hulɗa da juna kuma suka samar da wani fili na hasashen da ya dace.
Sakamako: Marubutan sun ba da rahoton cewa samfurin kirtani da yawa ya fi dabarun bin yanayi da kashi 15-20% dangane da rabon Sharpe a cikin gwajin baya na shekaru 2 akan bayanan EUR/USD. Algorithm ɗin ya kasance mai kwanciyar hankali ko da lokacin da aka ƙara farashin ciniki da kashi 50%.
8. Tsarin Bincike: Nazarin Shari'a
Nazarin Shari'a: Hasashen EUR/USD (2013-2015)
Mataki na 1: Tattara bayanan tick marasa daidaito daga OANDA.
Mataki na 2: Tsinkayi bayanan akan wani nau'in kirtani ta amfani da hanyar da aka bayyana a Sashe na 6.
Mataki na 3: Gano kirtani uku masu rinjaye masu dacewa da tsarin hauhawa, saukowa, da kuma gefe.
Mataki na 4: Yi amfani da tsarin kirtani da yawa don samar da alamun siye/sayarwa.
Mataki na 5: Gwada baya tare da farashin ciniki na 0.1% a kowace ciniki.
Sakamako: Riba mai tsafta na 12.3% a cikin shekaru 2, tare da matsakaicin raguwa na 8.1%.
9. Aikace-aikace na Gaba da Ra'ayi
Hanyar samfurin kirtani za a iya faɗaɗa ta zuwa wasu nau'ikan kadarori (kayayyaki, hannun jari) kuma a haɗa ta da koyon zurfi don cire siffofi. Bincike na gaba ya kamata ya mayar da hankali kan:
- Tushen lissafi mai ƙarfi na tsarin kirtani
- Haɗin kai tare da koyon ƙarfafawa don inganta fayil mai ƙarfi
- Aikace-aikace ga ciniki mai sauri tare da bayanan microsecond
- Kwatanta da hanyoyin nazarin bayanan topological kamar homology mai dorewa
10. Bincike na Asali (Kalmomi 300-600)
Wannan takarda ta Pinčák da Bartoš tana wakiltar wani ƙoƙari mai ƙarfi amma mai aibi na haɗa ka'idar kirtani da hasashen kuɗi. Babban ra'ayin—cewa kasuwannin kuɗi suna da tsarin geometric na asali—ba sabon abu ba ne; yana maimaita aikin farko akan geometry na faduwar kasuwa (Sornette, 2003) da amfani da geometry na Riemann a ka'idar fayil (Bielecki & Rutkowski, 2004). Duk da haka, takamaiman aikace-aikacen tsarin kirtani sabon abu ne kuma mai yuwuwar kawo canji.
Marubutan sun gano iyakokin samfuran tattalin arziki na gargajiya daidai: suna yin watsi da farashin ciniki, suna ɗaukan bayanan daidai, kuma sun kasa kama yanayin rashin Gaussian na dawowa. Waɗannan batutuwa ne da aka rubuta sosai (Cont, 2001; Mandelbrot, 1963). Samfurin kirtani yana ba da hanya don haɗa waɗannan siffofi ta hanyar ɗaukar hanyar farashin musaya a matsayin wani nau'i mai ƙarfi maimakon tsari mai yuwuwa.
Duk da haka, takardar tana fama da wasu raunuka masu mahimmanci. Na farko, tsarin lissafi bai cika ba. Marubutan ba su bayar da ma'anar 'tsarin kirtani' ba ko kuma yadda ya bambanta da dabarun koyon nau'i na yanzu (misali, Isomap, t-SNE). Na biyu, tabbatarwa ta zahiri ba ta da ƙarfi: an gwada kasuwa ɗaya (OANDA) da nau'in kuɗi ɗaya (EUR/USD) kawai, kuma lokacin gwajin baya shekaru biyu ne kawai. Na uku, kwatanta da dabarun bin yanayi bai isa ba; samfuran koyon injin na zamani (LSTM, Transformer) sun nuna aiki mafi girma a hasashen jerin lokuta (Sezer et al., 2020).
Daga mahangar aiki, bayanan masu aiki suna da iyaka. Takardar ba ta ba da algorithm da za a iya amfani da shi kai tsaye ko kuma bayyanannun jagororin aiwatarwa ba. Manufar 'tsarin kirtani da yawa' ta kasance mai ƙima, kuma ba a tattauna rikitarwar lissafi ba.
A ƙarshe, yayin da takardar ta buɗe wata hanya mai ban sha'awa don bincike, tana da nisa daga mafita mai aiki. Aikin gaba ya kamata ya mayar da hankali kan tushen lissafi mai ƙarfi, tabbatarwa ta zahiri mai yawa a kan kasuwanni da yawa, da kwatanta da samfuran zamani. Samfurin kirtani na iya samun wurinsa a cikin akwatin kayan aikin manazarta ƙididdiga, amma bai riga ya shirya don lokacin farko ba.
11. Manazarta
- Pinčák, R., & Bartoš, E. (2015). With string model to time series forecasting. arXiv:1511.00483.
- Sornette, D. (2003). Why Stock Markets Crash: Critical Events in Complex Financial Systems. Princeton University Press.
- Bielecki, T. R., & Rutkowski, M. (2004). Credit Risk: Modeling, Valuation and Hedging. Springer.
- Cont, R. (2001). Empirical properties of asset returns: stylized facts and statistical issues. Quantitative Finance, 1(2), 223-236.
- Mandelbrot, B. (1963). The variation of certain speculative prices. Journal of Business, 36(4), 394-419.
- Sezer, O. B., Gudelek, M. U., & Ozbayoglu, A. M. (2020). Financial time series forecasting with deep learning: A systematic literature review. Applied Soft Computing, 90, 106181.