Сборник тезисов «Ипак ва зираворлар»

Table 1. 
Data on the arrival of tourists and on income from tourism for 1997. 
Region 
Arrival, mln.h./Incoming, billion 
$. [15] 
1997 г. 
1998 г. 
America 
118,481 
120,190 
118,770 
121,230 
East Asia and the 
Pacific 
87,953 
86,930 
76,630 
73,740 
Near East 
14,833 
15,620 
9,140 
9,720 
Africa 
23,157 
24,900 
9,020 
9,550 
South Asia 
4,830 
5,070 
4,280 
4,400 
The whole world 
610,763 
625,24 
435,980 
444,740 
Table 2 shows the results of calculations for the number of tourists arrived in 
different regions. It can be seen that, in general, both the regression equations and the 
models analyzed have good predictive qualities. In this case, one should distinguish 
short-term forecasts for one period (one year) and medium-term forecasts (for two 
years). 
For short-term forecasts, the Holt and Brown models provide more accurate 
results. The smallest range of variation of values is provided by Brown's adaptive 
smoothing model, which is peculiar to "focus" on the trend. Less accurate results are 
obtained by regression equations. But, nevertheless, they can be fully used as a first 
approximation in forecasting. 
This conclusion leads to an analysis of medium-term forecasts, where the 
model of adaptive smoothing also gives the most accurate results. A characteristic 
feature of adaptive models is the presence of both overestimated estimates (with the 
"-" sign in the table) and undervalued ones, whereas for regression equations, as a 
rule, estimates are overestimated. This is explained by the effect of the factor of 
choosing the shape of the curve in the case of analytic alignment. 

---

365 
Summarizing the results of the research on the application of adaptive methods 
of mathematical modeling in tourism, it can be asserted that a number of results have 
been obtained, among which I would like to note the following: 
1. A promising trend in the modeling of processes in tourism, along with 
diffusion and econometric models is the use of models describing a linearly additive 
trend, i.e. Holt's model and Brown's additive smoothing model. 
2. It is confirmed that regression equations allow obtaining correct forecast 
estimates in tourism (the error, as a rule, does not exceed 20%). At the same time, 
there is some overestimation. 
3. It is established that the most accurate of the methods considered is the 
adaptive smoothing model. It allows to obtain correct prognostic estimates, the error 
of which does not exceed, as a rule, 10%. This model is not characterized by 
unilateral (constant) overstating or underestimation of estimates. 
4. The unevenness of regional economic processes in tourism does not have a 
significant effect on the nature of the use of adaptive models. 
5. When analyzing data from different regions, there is no need to perform 
additional assessment and adjustment of model parameters, which confirms their 
general nature. 
6. The issue of forecasting revenues from tourism requires additional study, 
because The models considered allow us to obtain only approximate estimates.