Pricing financial call options with a multilayer perceptron class of artificial neural network : case: S&P 500 index options in 2017-2019 ; Osto-optioiden hinnoittelu MLP-neuroverkolla : case: S&P 500- indeksin osto-optiot vuosina 2017-2019
2019
Hochschulschrift
Zugriff:
The objective of this Master’s thesis is to examine if a Multilayer Perceptron class of artificial neural network can be applied to estimate European call option prices on the S&P 500 index in 2017 to 2019. The estimations are also done with the Black-Scholes Model to benchmark the Multilayer Perceptron’s results. Two Multilayer Perceptron’s are built with the same architecture but with different activation functions. Furthermore, the options are partitioned by moneyness and maturity to further study the Multilayer Perceptron’s and Black-Scholes Model’s discrepancies. The final dataset constituted of 375 117 observations. The dataset was partitioned to training (70%) and test (30%) datasets. Rectified Linear Unit and Exponential Linear Unit activation functions were applied to the Multilayer Perceptron. The results are in line with previous research and confirm that a Multilayer Perceptron class of artificial neural network can be applied on European call option prices on the S&P 500 index in 2017-2019. The results of this thesis suggest that a Multilayer Perceptron estimates European call option prices on the S&P 500 index in 2017-2019 better than the Black-Scholes Model. Further, the results confirm that the Multilayer Perceptron can estimate European call option prices on the S&P 500 index in 2017-2019 better than the Black-Scholes Model despite the option’s moneyness or maturity. Exponential Linear Unit as an activation function is found to estimate better than the Rectified Linear Unit activation function. ; Tämän pro-gradu tutkielman tarkoituksena on selvittää, voiko MLP-neuroverkko estimoida S&P 500 -indeksin eurooppalaisia osto-optioiden hintoja vuosina 2017-2019. Optioiden hinnat estimoidaan vertailun vuoksi myös Black-Scholes-mallilla. Myös kaksi MLP-neuroverkkoa rakennetaan samalla rakenteella, mutta eri aktivointifunktioilla vertailun vuoksi. MLP-neuroverkon ja Black-Scholes-mallin tuloksia tutkitaan myös optioiden arvon (moneyness) ja maturiteetin mukaan. Lopullinen data sisälsi ...
Titel: |
Pricing financial call options with a multilayer perceptron class of artificial neural network : case: S&P 500 index options in 2017-2019 ; Osto-optioiden hinnoittelu MLP-neuroverkolla : case: S&P 500- indeksin osto-optiot vuosina 2017-2019
|
---|---|
Autor/in / Beteiligte Person: | Taillon, Sébastien ; Lappeenrannan-Lahden teknillinen yliopisto, LUT ; Lappeenranta-Lahti University of Technology, LUT |
Link: | |
Veröffentlichung: | 2019 |
Medientyp: | Hochschulschrift |
Schlagwort: |
|
Sonstiges: |
|