1. Indonesian Stock Market Characteristics based on Fractal Market Hypothesis by Gillian Tampi

Keywords: Fractal Market Hypothesis, Rescaled Range Analysis, Hurst Exponent, Brownian Motion.

This thesis aims to analyze the so-called fractal characteristic in Indonesian Stock Market using Rescaled Range Analysis with LQ45 Index, Jakarta Composite Index (IHSG), and five most active stocks for the last 16 years are used as the source data. The value of Hurst exponent yielded from Rescaled Range analysis was used to observe the return dynamics characteristics, measure correlation, calculate fractal dimension and test the significance of the Hurst exponent itself.

2. Optimal Hedging Strategy for Corn Consumption of Feed Mills in Indonesia by Michelle Sugita Gunawan

Keywords: bivariate garch, offshore commodity hedge , commodity future.

This thesis aims to estimate the optimal hedge ratio for corn futures and assess the effectiveness of offshore hedging strategy. It is found that ARCH effect is not present in the monthly and weekly series, thus OLS hedge ratio estimate is used.

3. Model Prediksi Kebangkrutan di Indonesia dengan Menggunakan *Support Vector Machine *oleh Kevin Yumantra

Keywords: *support vector machines, financial ratio, *prediksi probabilitas kebangkrutan.

In this thesis, a test was performed to examine the use of Support Vector Machine (SVMs). In this thesis SVMs will be used to predict the bankruptcy probability of the Indonesian companies based of the use of 14 financial ratio from those companies.

4. Prediction of ICU Patients’ Survival Using Survival Model and Its Comparison with SAPS II by Raden Mas Panji Widiarda R.

Keywords: survival model, cox proportional hazard, ICU, SAPS II, Baseline Survival Function, Cox and oakes estimator.

As an alternative to SAPS II, a scoring system used to calculate mortality of ICU patients, this thesis will discuss the construction of a survival function to predict ICU patients’ survivak using data from ICU in RSCM. Cox Proportional Hazard (PH) is chosen as the model, and the accuracy of this model will be tested using ROC analysis.

5. Modelling of Human Face and Expression Recognition Using Principal Component Analysis and Radial Base Function by Andreas Widijargo

Keywords: Face Recognition, Expression Recognition, Principal Component Analysis, Radial Basis Function.

Face recognition is one of the fields in pattern recognition that has been used for identification process. This research will discuss about human face recognition using Principal Component Analysis (PCA) and Radial Basis Function (RBF). For research expansion, a system of human expression recognition will be developed with the same models.

6. Image Restoration Using BSBC Model and Anisotropic Diffusion with Its Implementation Using Finite Difference Method by Yoel Martinus P.

Keywords: image inpainting, anisotropic diffusion, image restoration.

Image restoration is performed to remove noise and unwanted disturbance from an image. In this thesis, two equations are used to perform image restoration. The first equation is introduced by Bertalmio and friends which serves to draw back the broken part. The other equation is introduced by Perona and Malik which will be used in this thesis to smooth the result we get by performing the first equation.

7. Analysis and Simulation of Small-World and Scale-free Network by Michael Kam

Keywords: graph-theory, small-world, scale-free, nodes isolation.

Real network that are complex, not ordered, and yet not fully random graph, can be modeled with two approaches which are called small-world and scale-free. Small-world can be developed using edges rewiring, edges adding, and hotspots adding. On the other hand, scale-free network is built using growth and preferential attachment.

8. TimesSeries Forecasting Using Hybrid ARIMA and Artificial Neural Network Model by Andre Jonathan Cahyadi

Keywords: Time series forecasting, ARIMA model, Artificial Neural Network, conditional likelihood estimation, time series analysis.

ARIMA model is a linear model thus its capability in understanding the data characteristics is limited. In order to overcome this problem, other techniques have been developed, i.e. Artificial Neural Network (ANN). The hybrid approach combining both ARIMA and ANN will be used to increase forecasting accuracy.

9. Valuation of Currency Option Using the Hyperbolic Distribution Assumption by Cantika Puteri Utama

Keywords: currency option valuation, normal distribution, hyperbolic distribution, hyperbolic Levy motion, Black-Scholes formula.

According to the previous research, stock return can be expressed more precisely with the hyperbolic distribution assumption than the normal assumptions in the Black-Scholes formula. It also shows that the option prices under the hyperbolic distribution assumption are more similar to the market option price if they are compared to the option prices using the normal distribution assumption in the Black-Scholes formula. So, this study will analyze the option with currency as its underlying asset.

10. Numerical Simulation of Blood Flow In Atherosclerosis Condition Using Smoothed Particle Hydrodynamics Method by Yohanes Tjandrawidjaja

Keywords: atherosclerosis, blood flow, smoothed particle hydrodynamics, Navier-Stokes equations, inlet-outlet domain.

In order to have a better comprehension on the blood flow, one of the easiest way is by developing a numerical simulation. In this thesis, the numerical simulation is built upon the Smoothed Particle Hydrodynamics (SPH) method. The behavior of the blood flow is then observed by applying SPH to the Navier-Stokes Equations.

11. Multi-objective Optimization Using the Hybridization of Genetics Algorithm and Artificial Immune System by Ellen Pramana

Keywords: Multi-objective optimization, 0/1 knapsack problem, Genetics Algorithm, Artificial Immune System.

The classical optimization methods are limited to solve multi-objective optimization problem, thus many non classical methods, such as Genetics Algorithm and Artificial Immune System, is being developed. In this thesis, the hybrid approach of both of Genetics and Artificial Immune System will be used for solving the 0/1 knapsack problem.