CC BY
43
УДК 338.48
ДЖ. ТОНИНИ
РОССИЙСКИЙ ТУРИЗМ В ИТАЛИИ: ТЕНДЕНЦИИ, СЕЗОННОСТЬ, ОСОБЕННОСТИ
Российский туризм в Италии, несмотря на события и проблемы последних лет, продолжает оставаться важным и значимым для экономики Италии и, по прогнозам, будет укреплять свои позиции и развиваться в будущем. Но для данного туристского сегмента в настоящее время отсутствуют серьезные системные исследования, использующие соответствующие статистические модели с краткосрочными и долгосрочными временными компонентами.
В данной статье предпринимается попытка заполнить отмеченные пробелы в исследованиях, рассматриваются тенденции в развитии российского туризма, анализируется сезонность, в частности посредством анализа ночевок в период с января 2001 г по декабрь 2012 г. В исследовании использовались различные подходы и методы: метод декомпозиции, динамические модели множественной регрессии и др. Данные подходы позволят проанализировать феномен с различных точек зрения, а затем провести системный анализ полученных результатов.
Предложенные подходы будут использованы не только в отношении общего количества российских туристов, прибывающих в Италию, но и в отношении отдельных категорий данных туристов, выделяемых сектором гостеприимства (отелями и другими средствами размещения) во временные интервалы 2001-2006 гг. и 2007-2012 гг. Таким образом, посредством интерсекторального и интертемпорального сравнений можно лучше раскрыть характеристики российских туристских потоков в Италию. Данные характеристики будут в дальнейшем проанализированы на основе сопоставления ночевок российских туристов с ночевками туристов из других стран (в частности, немцев, итальянцев и др.)
Ключевые слова: российский туризм в Италии, туристский сегмент, сезонность, тенденции, прогнозы, сектор гостеприимства, ночевки, метод декомпозиции, динамические модели множественной регрессии, интерсекторальные и интертемпоральные сравнения.
In economic terms and otherwise, Russian tourism in Italy is becoming increasingly important; despite various historically contingent and adverse events, it will tend to consolidate and develop further in the future. However, despite the importance of this tourism segment, systematic studies that analyse, with appropriate statistical models, its short- and long-term temporal components are missing. This paper aims to fill this gap by developing a trend and seasonality analysis of Russian overnight stays in Italy from January 2001 to December 2012.
Two different approaches will be followed: the first will use a decomposition method based on appropriate moving averages, while the second will be based on dynamic multiple regression models with- out and with outlier detection.
This approach will enable to examine the phenomenon from different perspectives and then to proceed to a systematic comparison of the results obtained. These approaches will be applied not only to the total number of Russian tourists coming to Italy but also to its appropriate subsets, distinguished by hospitality sector (hotels and other tourist accommodations) and time interval (2001-2006 and 2007-2012). In this way, through appropriate intersectoral and intertemporal comparisons, the characteristics of Russian tourist flows to Italy can be better highlighted. These
© Тонини Дж., 2018
characteristics will be further analysed, comparing them with the overnight stays of two important tourist nationalities (German and Italian) as well as of all foreigners.
Keywords: Russian tourism in Italy; trends and seasonality; method based on moving averages; dynamic multiple regression model; outliers; intersectoral, intertemporal and international comparisons.
PART 1. TRENDS AND PATTERNS OF DEMAND AMONG RUSSIANS IN ITALY
1.1. Structural aspects of and trends in Russian tourist demand in Italy
This paper examines Russian tourist demand in Italy as one of the most important and dynamic segments of foreign tourism in this country. This demand will be considered in physical terms of overnight stays; in particular, the temporal components of trend and seasonality will be analysed. In both cases, the monthly time series of the tourist overnight stays recorded from January 2001 to December 2012 will be used and, in the case of the trend, the annual series of the period 2001-2012 will also be used (where 2012 is the last year with complete final data at the time of processing)1.
The analysis of trend and seasonality will be based on two different approaches, the first of which consists of the decomposition of the monthly series into short- and long-term components that were estimated using appropriate moving averages; in contrast, the second approach consists of the representation of the different components of a series via a dynamic multiple regression model.
Before moving to the analysis of the trend and seasonality of the monthly se-
1 Data on tourist overnight stays used in this study are all provided by ISTAT and are drawn from: ISTAT (2002 and 2004) for the years 2001 and 2002; ISTAT (2005-2011) for the years 2003-2010; ISTAT (2012-2014) for the years 2011-2012. The terms «overnight stays», «overnights», «nights spent» and «nights accomodation» will be used as synonyms.
ries under examination, it is useful to analyse some important features that can be gleaned from the corresponding annual series relating to the 2001-2012 interval. In this con- text, Table 1.1 shows the annual nights spent by Russian tourists, broken down by hotels and other tourist accommodations; moreover, for the purposes of comparison with other nationalities, the table also contains the total overnight stays of German2, foreign3 and Italian tourists.
An analysis of column (3) of Table 1.1 suggests that the total number of overnights spent by Russian tourists is still a minority segment of foreign tourism in Italy, at least compared with Germans, who represent approximately 30% of foreign nights spent. However, during the 12 years in question, the weight of the total overnight stays of Russian tourists increased significantly from 0.92% to 3.42% of total foreign tourist overnights. On the contrary, the share of German nights spent decreased considerably, dropping from 36.60% in 2001 to 28.66% in 2012. Similarly, even
2 In the case of German tourists, reference is made to tourists coming from Germany, which is the main country of origin of foreign tourists staying in Italy, and not to German-speaking tourists. In the following, therefore, «German» is to be understood as «Germany resident».
3 The number of foreign overnights under consideration relates to the overnight stays of all foreign tourists who have stayed in Italy, including Germans. This choice allows appropriate comparisons with the complementary subset of «non-foreign» tourists, that is Italians, reported by ISTAT.
the percentage of Italian overnights decreased, moving from 58.13% to 52.56% between 2001 and 2012. Conversely, the proportion of the foreign nights spent increased, also thanks to the contribution
of the growth of Russian tourists. In summary, while traditional foreign segments, such as the German one, lost importance, the weight of new segments, such as the Russian one, increased.
Table 1.1
Russian overnight stays Total overnight stays by
YEAR Hotels (1) Other tourist accommodat. (2) Totals (3) Germans (4) Foreigners (5) Italians (6)
2001 1,226 117 1,343 53,687 146,672 203,651
2002 1,204 126 1,330 50,810 145,560 199,687
2003 1,282 143 1,426 46,503 139,653 204,760
2004 1,488 147 1,636 45,184 141,165 204,151
2005 1,781 162 1,943 44,441 148,501 206,754
2006 2,377 178 2,555 46,401 156,861 209,903
2007 3,154 252 3,406 46,497 163,466 213,176
2008 3,441 290 3,730 45,402 161,797 211,869
2009 2,966 329 3,295 47,278 159,494 211,269
2010 3,802 407 4,209 47,802 165,202 210,340
2011 4,753 566 5,319 50,200 176,474 210,421
2012 5,502 678 6,180 51,752 180,595 200,116
(1)/(3)0/o (2)/(3)% (3)/(5)00 (4)/(5)00 (5)/[(5)+(6)]/ (6)/[(5)+(6)]/
2001 91.26 8.74 0.92 36.60 41.87 58.13
2002 90.52 9.48 0.91 34.91 42.16 57.84
2003 89.94 10.06 1.02 33.30 40.55 59.45
2004 91.00 9.00 1.16 32.01 40.88 59.12
2005 91.64 8.36 1.31 29.93 41.80 58.20
2006 93.02 6.98 1.63 29.58 42.77 57.23
2007 92.61 7.39 2.08 28.44 43.40 56.60
2008 92.23 7.77 2.31 28.06 43.30 56.70
2009 90.00 10.00 2.07 29.64 43.02 56.98
2010 90.33 9.67 2.55 28.94 43.99 56.01
2011 89.36 10.64 3.01 28.45 45.61 54.39
2012 89.03 10.97 3.42 28.66 47.44 52.56
Annual series 2001-2012 of Russian overnight stays (hotels, other tourist accommodations, totals) and of German, foreign and Italian overnight stays (values in thousands). Percentages, variation ratios (VRs), and continuous growth rates (CGRs)
VRs 2001-12 +389 +51 +440 -176 +3,084 -321
VRs 2001-06 +230 + 12 +242 -1,457 +2,038 + 1,250
VRs 2007-12 +470 +85 +555 + 1,051 +3,426 -2,612
CGRs 2001-12 + 13.65% + 15.97% +13.88% -0.33% +1.89% -0.16%
CGRs 2001-06 + 13.24% +8.39% +12.86% -2.92% +1.34% +0.60%
CGRs 2007-12 + 11.13% + 19.79% +11.92% +2.14% +1.99% -1.26%
Source: Our processing of ISTAT data (2002 and 2004), (2005-2011) and (2012-2014).
The latter segment is characterised by the strong preference given to hotels, the site of approximately 90% of Russian overnight stays. The weight of hotels, however, de- creased slightly over time, dropping from 91.26% in 2001 to 89.03% in 2012. In front of this decrease, there is a corresponding increase in the proportion of nights spent in other tourist accommodations, even if, in the 12 years 2001-2012, the percentage of Russian overnights in these accommodations never exceeded 11%.
Passing from percentages to variation ratios1, it is possible to grasp more accurately the evolution of Russian tourism in Italy during the period 2001-2012 and sub-6 periods 2001-2006 and 2007-20122.
1 Variation ratios are expressed as VR = (Pt+k -Pt) / k, where Pt and Pt+k indicate the overnights respectively of the initial year t and the final year t+k, with t = 2001 and k = 11 in the case of the period 2001-2012; t = 2001 and k = 5 in the case of the 2001-2006 interval; and, finally, t = 2007 and k = 5 in the case of the 2007-2012 interval.
2 The division into two sub-periods of the 2001-
2012 interval is justified by the hypothesis that, in the course of the 12 years in consideration, the trend of tourists under examination is not homogeneous but presents a discontinuity between the first and the second half of the 2001-2012 interval.
In particular, the average annual variation of Russian overnights is always positive and equal to 440,000 more nights spent per year in 7the 12 years under consideration, which corresponds to a continuous growth rate of 13.88%3. The most sustained annual average increments are recorded in the period 2007-2012, with over 555,000 more overnight stays per year compared with 242,000 in the period 2001-2006.
Among the segments considered, in addition to the Russian overnight stays, only the foreign tourists provide variations that are always positive, even if the continuous growth rate is markedly lower than that of Russians, never exceeding 2%. Conversely, the sign of the average annual variations of German and Italian nights spent change over time, although in both cases they are negative in the 12 years 2001-2012. Specifically, German overnights decrease by almost 1,460,000 units per year in the period 2001-2006, while they increase by more than 1,050,000 units per year in the
3 The continuous or exponential growth rate is expressed by CGR = ln(Pt+k / Pt) / k • 100, where Pt and Pt+k indicate the overnight stays of the initial year t and the final year t+k, respectively, which, depending on the cases, correspond to the years 2001 and 2012, 2001 and 2006, or 2007 and 2012.
next interval 2007-2012 without, however, reaching the level of 2001. In contrast, the nights spent by Italian tourists increase by 1,250,000 units per year in the period 2001-2006, while they decrease by more than 2,600,000 units per year in the next interval, thus falling below the level of 2001.
The values of the indices reported in Table 2.1 confirm the loss of importance of traditional tourist segments, such as the German and Italian ones, and highlight the emergence of new foreign segments, including the Russian one, characterised by a significant expansion over time and by high levels of spending.
1.2. Trends in Russian overnight stays estimated from annual and monthly time series
The very strong growth of Russian overnight stays, documented through the high values of the relative continuous growth rate (Tab. 1.1) and through the graphs in Figure 1.1, highlights a non-linear trend during the period 2001-2012. To reproduce this trend, it was decided, among the various functions tested, to opt for a quadratic function, whose parabolic arc is concave upward. This non-linear function interpolates the observed values well not only with reference to the 12 years 2001-2012 but also for each of the two sub-periods 2001-2006 and 2007-2012 (Fig. 1.1). The only exception is related to the nights spent by Russians in other tourist accommodations in 2001-2006, for which a linear interpolation seems more appropriate (not shown for reasons of space).
The trend of German and Italian total overnight stays is represented well by the parabola, even if in the first case it is concave upward, while in the second
downward (Fig. 1.1). The reason for this outcome is, of course, that for German overnights, the first period of decline is followed by a subsequent recovery; conversely, Italian over- nights are characterised by an initial growth followed by a noticeable decline. Finally, the low growth rate of foreign total nights spent argues in favour of a linear trend, which is also confirmed by the fluctuations of the observed values around the best fit line.
The trend was estimated not only on annual series but also on monthly series, as shown in Figure 1.2. In the latter case, the trend was initially estimated via a centred moving average; subsequently, the initial estimate of the trend was interpolated with a linear or quadratic function depending on the shape of the values of the moving aver- age.
Predictably, the final estimate of the monthly trend has led to results in line with those obtained previously. In particular, the quadratic function has shown a good fit to the values of the moving average with reference to almost all of the monthly series considered. The only exceptions are the Russian overnight stays in other tourist accommodations which, from January 2001 to December 2006, show a linear trend, as do the foreign total nights spent, characterised by linearity at least since January 2004.
1.3. Seasonality of Russian overnight stays: Characteristics and comparisons
A seasonality analysis of tourist overnights will be based, initially, on monthly seasonal indices obtained by using appropriate moving averages. The values of these indices are shown in Table 1.2 with reference to Russian nights spent in hotels, in other and in total tourist
Russian overnights in hotels 2001-2012
Non-hotel Russian overnights 2001-2012
Russian total overnight stays 2001-2006
Russian total overnight stays 2007-2012
Russian total overnight stays 2001-2012
German total overnight stays 2001-2012
Foreign total overnight stays 2001-2012
Italian total overnight stays 2001-2012
Figure 1.11.1 Russian overnight stays in hotels, in other (non-hotel) and in total tourist accommodations; total nights spent by German, foreign and Italian tourists. Observed and interpolated series (smooth curve) for the years 2001-2012, 2001-2006 and 2007-2012
1 The source of all the figures is the processing of ISTAT data (2002 and 2004), (2005-2011) and (2012-2014).
Russian overnights in hotels 2001-2012
rrïiTïïïwmîîiâiTïTïïwmî"
Russian total overnight stays 2001-2012
wîïiïïîmiiîTîTmwnîP ü g üi s-ftí s i g Mi i ü il №ï ë s»ts
Non-hotel Russian overnights 2001-2012
"1ÏWÎPÎPÏÎÎWÏÎTÎMÏÎÎPÎPIÏÎ3™
German total overnight stays 2001-2012
TïîïïîsTïîiriîîïïïWïTïîïïîîïîïr imSimSBIiífmmijmmSH
Russian total overnight stays 2001-2006
siiiiNeilpiiiitiigii^iiiMHiiiiiei
Foreign total overnight stays 2001-2012
TïïïirrmiTïïïïïîTïTïïïîïîiT iisiiiiiliiïSBgâiaïisisiïiip
Russian total overnight stays 2007-2012
irmrrrrîTTiiTîTïTTiTr.
II fä S S I ä S ä ! ä N a a a s a g s s
Italian total overnight stays 2001-2012
í î s i 4 s ü st mim* s í itat|í»í
IHiillf?*ÍK>J§{ iäiiäüiniss
Figure 1.2. Monthly overnight stays of Russian, German, foreign and Italian tourists, and
corresponding values of the trend
accommodations for the period January 2001-December 2012, as well as to total Russian nights spent in the sub-periods January 2001-December 2006 and January 2007 - December 2012.
By way of comparison, finally, the table also contains the indices of total German, foreign and Italian overnight stays during the interval January 2001-December 20121. In the period 2001-2012, the seasonal indices of total Russian overnight stays exceeds level 100 of monthly equi-distribution only in the summer period from June to September, peaking in August, followed by July and September, which overtake June (Tab. 2.2). The value for the months of January and May is close to 100, while the indices for all other months are well below the level of equipartition and reach a minimum in December.
In the period 2001-2012, the seasonal indices of total Russian overnight stays exceeds level 100 of monthly equidistribu-tion only in the summer period from June to September, peaking in August, followed by July and September, which overtake June (Tab. 2.2). The value for the months of January and May is close to 100, while the indices for all other months are well below the level of equipartition and reach a minimum in December.
In the period January 2001-December 2012, Russian nights spent in hotels show values of seasonal indices similar, month by month, to those observed above for the Russian overnight stays in total tourist accommodations; this is obviously because
1 Seasonal indices are expressed, for convenience, in hundreds rather than in unit terms. Therefore, the value of equipartition or uniform seasonality is equal to 100 and the total to 1,200. Consequently, months with values of indices greater than 100 are those in which are concentrated, more or less intensely, the overnight stays. The opposite obviously occurs for months with seasonal indices less than 100.
the hotel overnights amount to approximately 90% of the total overnights. Moving from the hotel sector to other tourist accommodations, the concentration of overnight stays in the months June-September increases; in these four months, in fact, it is concentrated in more than 64% of annual overnight stays in other tourist accommodations, compared with just over 50% in hotels and in total tourist accommodations (Tab. 2.2 and Fig. 2.3). This result confirms the higher seasonal concentration that usually characterises other tourist accommodation facilities compared with hotel facilities, also because, in the hotels, the seasonal closures of these establishments are generally less frequent and shorter.
Russian tourism, therefore, is characterised as summer vacation tourism primarily focussed on the seacoast, especially in Emilia-Romagna, and as visitation to the main Italian art cities (Venice, Rome and Florence in the lead, with the addition of Milan as a city of business, fashion and shopping). Significant, although less numerous, are the flows registered in January and May, corresponding, in the first case, with the Ortho- dox Christmas, winter tourism and the end-of-season sales and, in the second case, with important Russian public holidays.
Over time, there have been noticeable changes in the seasonality of Russian total overnight stays, and these changes have led to a higher monthly concentration of Russian tourist demand. In fact, between 2001-2006 and 2007-2012, the concentration of overnight stays in the five months of higher inflows increased from 58.85% to 63.04%; moreover, the maximum value of the month of August rose from 173.03 to 186.04, while the December minimum fell from 48.58 to 44.19, leading to a significant broadening of the range of variation (increasing from 124.45 to 141.85).
Table 12
Monthly seasonal indices of Russian, German, foreign and Italian overnight stays. The values are expressed in hundreds and calculated using a moving average applied to the monthly series 2001-2012 and, for the Russian total overnights, also to the monthly
series 2001-2006 and 2007-2012
Russian overnight stays 2001-2012 Russian overnight stays in total tourist accommodations
Hotels Other structures 2001-2006 2007-2012
January 100.37 69.14 83.73 108.25
February 62.47 42.30 62.96 55.58
March 74.85 46.73 75.44 67.22
April 73.19 53.07 76.91 64.45
May 99.70 80.11 108.76 87.53
June 117.67 158.66 115.19 128.35
July 159.72 238.86 158.15 183.19
August 173.22 234.27 173.03 186.04
September 153.25 137.64 151.08 150.59
October 81.82 62.00 83.44 76.30
November 57.41 38.63 62.73 48.33
December 46.32 38.60 48.58 44.19
Overnight stays in total tourist accommodations 2001-2012
Russian German Foreign Italian
January 97.67 33.18 41.73 53.92
February 60.57 39.59 51.12 49.50
March 72.40 56.42 62.35 55.73
April 71.63 79.27 79.82 62.53
May 98.01 135.23 115.35 68.68
June 121.41 167.14 143.33 140.94
July 166.65 158.15 209.48 217.74
August 178.34 220.81 190.09 301.75
September 151.74 185.24 150.02 107.09
October 80.15 87.33 87.65 51.69
November 55.79 17.11 35.64 39.24
December 45.66 20.53 33.42 51.19
Source: Our processing of ISTAT c lata (2002 and 2004), (2005-2011) and (2012-2014).
Finally, between 2001-2006 and 2007- sitions (with August always in the lead,
2012, the ranking of the five months with followed by July and September, which
seasonal in- dices exceeding 100 does not overtake June), while May in the fifth posi-
experience variations in the first four po- tion is replaced by January.
Russian total overnight stays 2001-2012
Russian total overnight stays 2001-2006 (lower peak) and 2007-2012
t t M M M 1 I I I
Russian overnight stays in hotels (lower peak) and other tourist accommodations 2001-1012
M M M M M I I
Russian (lower peak) and German total overnight stays 2001-2012
I i ! I I | I ) | | | |
Russian (lower peak) and foreign total overnight stays 2001-2012
!MMMI|I||
Russian (lower peak) and Italian total overnight stays 2001-2012
MHm|i|i
Figure 2.3. Monthly seasonal indices of Russian, German, foreign and Italian total overnight stays in 2001-2012, and for the Russians, even in 2001-2006 and 2007-2012. The curves with the lowest maximum point represent the Russian overnight stays in 2001-2006, in the
hotel sector and totals, respectively
The increase in seasonality of Russian overnight stays in the 12 years 2001-2012 is also evident from Figure 2.3. In fact, comparing the variations in the seasonal indices of 2007-2012 with those of 20012006, it is found that the first exceeds the second in the summer months (excluding September), while exactly the opposite is the case at the tails (with the already-reported exception of January). The above-mentioned in- crease in seasonal concen-
tration is also confirmed by calculating the coefficient of variation of the seasonal indices. This coefficient shows a significant increase, from 39.49% to 48.79%, between 2001-2006 and 2007-2012.
During 2001-2012, the tendency of Russian tourists to choose summer holidays, spent mainly on the seacoast, increases; moreover, winter stays - related to the Orthodox Christmas, to skiing at Alpine resorts, and to the end-of-season sales in January - are
preferred to spring tourism, which is linked to important Russian public holidays in May.
To fully understand the characteristics of the seasonality of Russian overnight stays during the period 2001-2012, it is appropriate to compare it with the seasonality of the main foreign tourist segment, the German one, as well as with the seasonality of the foreign and Italian tourists (Tab. 2.2 and Fig. 2.3). This comparison suggests that, in relation to seasonal concentration, Italian overnight stays occupy the first position, followed by the German and foreign overnights, while Russian nights spent show a weaker seasonality. This ranking does not vary with changes in the measure of seasonality considered (average of seasonal indices higher than 100, maximum seasonal index and range of variation of seasonal indices),
with Italians always in the lead (191.88; 301.75; 262.51), the Germans in second position (173.31; 220.81; 203.70), the foreigners in third (161.65; 209.48; 176.06) and, finally, Russians in the fourth and last position (154.54; 178.34; 132.68).
In each case, August is the peak month not only for traditional segments of Italian and German tourists but also for the new segment of Russians who, at least in the case of overnight stays in hotels, are aligned, although to a lesser extent, with the behaviour of other nationalities. In contrast, for Russian overnights in other tourist accommodations and foreign total nights spent, the peak month is July followed by August, even if, in the case of the Russians, the distance between the two months is not very high.
■
I ЛИТЕРАТУРА и ЭЛЕКТРОННЫЕ РЕСУРСЫ
■
1. Abraham B, Chuang A. Outlier detection and time series modeling// Technometrics. 1989. Vol. 31, № 2. P. 241-248.
2. Akaike H. A bayesian extension of the minimum AIC procedure // Annals of the Institute of Statistical Mathematics. 1978. Vol. 30, A. P. 9-14.
3. Alivernini A., Breda E, Iannario E. International tourism in Italy (1997-2012)// Questioni di Economia e Finanza. 2014. №. 220, Banca d'Italia, Rome.
4. Alivernini A, Gallo M., Mattevi E, Quintiliani F. Tendenze del turismo internazionale nelle regioni italiane // Rassegna Economica. 2012. № 1.
5. Banca d'Italia. Turismo internazionale dell'Italia. Accessed December 2014. URL: https:// www.bancaditalia.it/statistiche/rapp_estero/ turismo-int
6. Banca d'Italia. Turismo internazionale - Dati. Accessed November 2015. URL: https://www. bancaditalia.it/statistiche/tematiche/rapporti-estero/turismo-internazionale/index.html
7. Banca d'Italia. Turismo internazionale - Dati. Accessed April 2016. URL: https://www.banca-ditalia.it/statistiche/tematiche/rapporti-estero/ turismo- internazionale/index.html
8. Box G. E.P., Jenkins G. M. Time Series Analysis, Forecasting and Control (rev. ed, 1976). Hold-en-Day, San Francisco, 1970.
9. Chang I., Tiao G. C, Chen C. Estimation of time series parameters in the presence of outliers // Technometrics. 1988. Vol. 30. P. 193-204.
10. Chen C., Liu L. Joint estimation of model parameters and outliers effects in time series // Journal of the American Statistical Association.
1993. Vol. 88. P. 284 -297.
11. Dagum E. B. Analisi delle serie storiche. Modellistica, previsione e scomposizione. SpringerVerlag Italia, Milan, 2002.
12. Enit Mosca. Mercato turistico russo. 2012.
13. Fox A.J. Outliers in time series // Journal of the Royal Statistical Society. 1972. Series B. Vol. 34. P. 350-363.
14. Furmanov K, Balaeva O, Predvoditeleva M. Tourism flows from the Russian Federation to the European Union// Anatolia: An International Journal of Tourism and Hospitality Research. 2012. Vol. 23, № 1. P. 17-31.
15. Gupta M, Gao J., Aggarwal C. C., Han J. Outlier detection for temporal data: A survey, IEEE Transactions on Knowledge and Data Engineering. 2014. Vol. 25, № 1. P. 1-20.
