Consumers have many options for shopping their daily needs, both in modern and traditional markets, or retail which especially minimarkets. Thus the competition among minimarket is also very high. Each minimarket has different marketing strategies, because the characteristics of consumers in shopping also vary. Marketing strategy need information from various aspects from consumers, competitors, and products are needed. This study analyze factors that influence the interest of consumers who shop at minimarket. Factor studied is in terms of consumers, both characteristics and location of residence. The data used is the primary data by conducting interviews survey on consumers in the Ngaglik District, Sleman Regency, Special Region of Yogyakarta. The minimarket sample is Indomaret. The method analysis are logistic regression and Classification and Regression Trees (CART). The result indicate that the significant factors that influence to shopping interest with logistic regression method are gender, monthly average expenditure, and location. While the factor plays an important role in CART is also the location. The CART, as a nonparametric method that doesn't have a certain distribution assumption, has higher classification accuracy. This is indicated by the percentage value of classification accuracy in CART is 90%, while logistic regression is 88%.

Gold investment has some benefits, it’s not only safe in value but also easily stored, durable, and the rate of return is relatively stable even several years tends to rise. Gold prices give us positive and negative side depending on the different event. To managing the negative side of holding gold, the investor has to measure the risk of gold at a given period. The most popular measurement of gold investment risk is Value at Risk (VaR). Determination of inappropriate methods will make VaR calculations inaccurate. The gold’s price from year to year is suspected to have fat tail distributed (heavy tail), the Extreme Value Theory (EVT) is considered as precise methods to find VaR. In this study, the Generalized Extreme Value (GEV) approach used in EVT estimated. GEV distribution identifies extreme values based on the maximum value of each block. Test results show that monthly block usage yield VaR value is more accurate than 0,899%. It means that for one coming period with a 95% confidence level the maximum loss that investors may experience is 0.899% of the total investment.

No abstract

No abstract

Value at Risk (VaR) is one of the risk measurement techniques and is considered a standard method of measuring risk. EWMA is one method to find standard deviation value of Conditional Variance which used to calculate the VaR. Investors use VaR to determine the risk level. VaR defined as the estimated loss of portfolio with a certain level of confidence. A portfolio composed of several mixed mutual funds. Of the four mutual funds only two mutual funds that can be arranged to get an optimum portfolio, 20% of mutual funds Kresna Flexima and 80% Nikko BUMN Plus. Portfolio VaR is calculated by EWMA method because it found the existence of conditional Variance. With a 95% level of trust and decay factor in accordance with the proposed risk metrics of 0.94 for daily data than obtained the VaR value of 0.26221011. This means that the maximum losses that may be received by investors amounted to 26.22% if investors invest in assets recommended by the optimal portfolio. This level of risk will be used by investors to control investment risk. Keyword: VaR, EWMA, Portofolio Optimum, Reksadana