Data analysis of personalized investment Decision Making using robo-advisers
Introduction. Nowadays, the problem of the optimal balance between consumption and savings, transformed into investments is solved by using automated systems for making investment decisions, such as robo-advice services which have the mathematical algorithm based on the main principles of consumptio...
Gespeichert in:
| Veröffentlicht in: | Наука та інновації |
|---|---|
| Datum: | 2020 |
| Hauptverfasser: | , , , |
| Format: | Artikel |
| Sprache: | English |
| Veröffentlicht: |
Видавничий дім "Академперіодика" НАН України
2020
|
| Schlagworte: | |
| Online Zugang: | https://nasplib.isofts.kiev.ua/handle/123456789/184838 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| Назва журналу: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Zitieren: | Data analysis of personalized investment Decision Making using robo-advisers / V. Kobets, V. Yatsenko, A. Mazur, M. Zubrii // Наука та інновації. — 2020. — Т. 16, № 2. — С. 87-100. — Бібліогр.: 25 назв. — англ. |
Institution
Digital Library of Periodicals of National Academy of Sciences of Ukraine| id |
nasplib_isofts_kiev_ua-123456789-184838 |
|---|---|
| record_format |
dspace |
| spelling |
Kobets, V. Yatsenko, V. Mazur, A. Zubrii, M. 2022-07-20T11:25:45Z 2022-07-20T11:25:45Z 2020 Data analysis of personalized investment Decision Making using robo-advisers / V. Kobets, V. Yatsenko, A. Mazur, M. Zubrii // Наука та інновації. — 2020. — Т. 16, № 2. — С. 87-100. — Бібліогр.: 25 назв. — англ. 1815-2066 DOI: doi.org/10.15407/scin16.02.087 https://nasplib.isofts.kiev.ua/handle/123456789/184838 Introduction. Nowadays, the problem of the optimal balance between consumption and savings, transformed into investments is solved by using automated systems for making investment decisions, such as robo-advice services which have the mathematical algorithm based on the main principles of consumption-savings theories. Problem Statement. The task assignment of developed IT service is to maintain a constant level of client’s consumption during life-long period through automated analysis of how much he/she has to consume and save each year. Results of consumption and savings proposals can be modified if initial financial data changes. Purpose. To develop investment plan of investors’ profiles taking into account their risk preferences using data analysis of robo-adviser service. Materials and Methods. SWOT-analysis of robo-advice (RA) services and comparative characteristics of robo-advisers explain advantage of RA services. Microservice for calculating stable consumption, finance consulting model of robo-advisor to ensure a constant level of consumption for the client are developed using the following technologies: Python 3.6, Django 2.0, Django Rest framework, AngularJs, HTML5, CSS 3, Bootstrap. Results. We considered consumption-saving ratio in economics, emerging trends of robo-advice (RA) services for making investment decisions. A mathematical model of robo-advisor in long-run period was developed and the support of investment decision making was described using micro-service of robo-advisor. Conclusions. The development RA is intended primarily for private persons (investors) who invest in long- term financial instruments in order to provide them with a permanent passive income based on their chosen savings period and the moment of retirement. Вступ. На сьогодні проблема пошуку оптимального балансу між споживанням та заощадженнями, перетвореними в інвестиції, вирішується за допомогою автоматизованих систем прийняття інвестиційних рішень, прикладом яких є послуги автоматизованих фінансових консультантів або робо-консультантів (robo-advisers), які працюють на базі математичного алгоритму, що ґрунтується на теорії споживання та заощадження. Проблематика. Завданням розробленого ІТ сервісу є підтримка постійного рівня споживання клієнта протягом всього періоду життя шляхом автоматичного аналізу того, скільки він/вона має споживати та заощаджувати щороку. Результати пропозицій щодо співвідношення споживання та заощадження можуть змінюватися при зміні початкових фінансових даних. Мета. Розробити інвестиційний план для профайлів інвесторів з урахуванням їх схильності до ризику за допомогою аналізу даних сервісом автоматизованих фінансових консультантів (robo-advisers). en Видавничий дім "Академперіодика" НАН України Наука та інновації Cвіт інновацій Data analysis of personalized investment Decision Making using robo-advisers Аналіз даних персоналізованого прийняття інвестиційних рішень з використанням автоматизованих фінансових консультантів Article published earlier |
| institution |
Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| collection |
DSpace DC |
| title |
Data analysis of personalized investment Decision Making using robo-advisers |
| spellingShingle |
Data analysis of personalized investment Decision Making using robo-advisers Kobets, V. Yatsenko, V. Mazur, A. Zubrii, M. Cвіт інновацій |
| title_short |
Data analysis of personalized investment Decision Making using robo-advisers |
| title_full |
Data analysis of personalized investment Decision Making using robo-advisers |
| title_fullStr |
Data analysis of personalized investment Decision Making using robo-advisers |
| title_full_unstemmed |
Data analysis of personalized investment Decision Making using robo-advisers |
| title_sort |
data analysis of personalized investment decision making using robo-advisers |
| author |
Kobets, V. Yatsenko, V. Mazur, A. Zubrii, M. |
| author_facet |
Kobets, V. Yatsenko, V. Mazur, A. Zubrii, M. |
| topic |
Cвіт інновацій |
| topic_facet |
Cвіт інновацій |
| publishDate |
2020 |
| language |
English |
| container_title |
Наука та інновації |
| publisher |
Видавничий дім "Академперіодика" НАН України |
| format |
Article |
| title_alt |
Аналіз даних персоналізованого прийняття інвестиційних рішень з використанням автоматизованих фінансових консультантів |
| description |
Introduction. Nowadays, the problem of the optimal balance between consumption and savings, transformed into investments is solved by using automated systems for making investment decisions, such as robo-advice services which have the mathematical algorithm based on the main principles of consumption-savings theories.
Problem Statement. The task assignment of developed IT service is to maintain a constant level of client’s consumption during life-long period through automated analysis of how much he/she has to consume and save each year. Results of consumption and savings proposals can be modified if initial financial data changes.
Purpose. To develop investment plan of investors’ profiles taking into account their risk preferences using data analysis of robo-adviser service.
Materials and Methods. SWOT-analysis of robo-advice (RA) services and comparative characteristics of robo-advisers explain advantage of RA services. Microservice for calculating stable consumption, finance consulting model of robo-advisor to ensure a constant level of consumption for the client are developed using the following technologies: Python 3.6, Django 2.0, Django Rest framework, AngularJs, HTML5, CSS 3, Bootstrap.
Results. We considered consumption-saving ratio in economics, emerging trends of robo-advice (RA) services for making investment decisions. A mathematical model of robo-advisor in long-run period was developed and the support of investment decision making was described using micro-service of robo-advisor.
Conclusions. The development RA is intended primarily for private persons (investors) who invest in long- term financial instruments in order to provide them with a permanent passive income based on their chosen savings period and the moment of retirement.
Вступ. На сьогодні проблема пошуку оптимального балансу між споживанням та заощадженнями, перетвореними в інвестиції, вирішується за допомогою автоматизованих систем прийняття інвестиційних рішень, прикладом яких є послуги автоматизованих фінансових консультантів або робо-консультантів (robo-advisers), які працюють на базі математичного алгоритму, що ґрунтується на теорії споживання та заощадження.
Проблематика. Завданням розробленого ІТ сервісу є підтримка постійного рівня споживання клієнта протягом всього періоду життя шляхом автоматичного аналізу того, скільки він/вона має споживати та заощаджувати щороку. Результати пропозицій щодо співвідношення споживання та заощадження можуть змінюватися при зміні початкових фінансових даних.
Мета. Розробити інвестиційний план для профайлів інвесторів з урахуванням їх схильності до ризику за допомогою аналізу даних сервісом автоматизованих фінансових консультантів (robo-advisers).
|
| issn |
1815-2066 |
| url |
https://nasplib.isofts.kiev.ua/handle/123456789/184838 |
| citation_txt |
Data analysis of personalized investment Decision Making using robo-advisers / V. Kobets, V. Yatsenko, A. Mazur, M. Zubrii // Наука та інновації. — 2020. — Т. 16, № 2. — С. 87-100. — Бібліогр.: 25 назв. — англ. |
| work_keys_str_mv |
AT kobetsv dataanalysisofpersonalizedinvestmentdecisionmakingusingroboadvisers AT yatsenkov dataanalysisofpersonalizedinvestmentdecisionmakingusingroboadvisers AT mazura dataanalysisofpersonalizedinvestmentdecisionmakingusingroboadvisers AT zubriim dataanalysisofpersonalizedinvestmentdecisionmakingusingroboadvisers AT kobetsv analízdanihpersonalízovanogopriinâttâínvesticíinihríšenʹzvikoristannâmavtomatizovanihfínansovihkonsulʹtantív AT yatsenkov analízdanihpersonalízovanogopriinâttâínvesticíinihríšenʹzvikoristannâmavtomatizovanihfínansovihkonsulʹtantív AT mazura analízdanihpersonalízovanogopriinâttâínvesticíinihríšenʹzvikoristannâmavtomatizovanihfínansovihkonsulʹtantív AT zubriim analízdanihpersonalízovanogopriinâttâínvesticíinihríšenʹzvikoristannâmavtomatizovanihfínansovihkonsulʹtantív |
| first_indexed |
2025-11-27T01:48:47Z |
| last_indexed |
2025-11-27T01:48:47Z |
| _version_ |
1850791897565495296 |
| fulltext |
87
СВІТ ІННОВАЦІй
https://doi.org/10.15407/scin16.02.087
koBEts, V.1, yatsEnko, V.2, MaZur, a.1, and ZuBrii, M.1
1 Kherson State University,
27, Universitetska St., Kherson, 73003, Ukraine,
+380 552 32 6782, kipiekedu@gmail.com
2 Taras Shevchenko National University of Kyiv,
90-A, Vasylkivska St., Kyiv, 03022, Ukraine,
+380 44 521 3578, decanat_econom@univ.kiev.ua
data anaLYSiS oF PeRSonaLiZed inveStMent
deciSion MaKinG USinG RoBo-adviSeRS
цитування: kobets, V., yatsenko, V., Mazur, a., and Zubrii, M. Data analysis of personalized investment
Decision Making using robo-advisers. Nauka innov. 2020. V. 16, no. 2. p. 87—100. https://doi.org/10.15407/
scin16.02.087
Introduction. Nowadays, the problem of the optimal balance between consumption and savings, transformed
into investments is solved by using automated systems for making investment decisions, such as roboadvice ser
vices which have the mathematical algorithm based on the main principles of consumptionsavings theories.
Problem Statement. The task assignment of developed IT service is to maintain a constant level of client’s
consumption during lifelong period through automated analysis of how much he/she has to consume and save
each year. Results of consumption and savings proposals can be modified if initial financial data changes.
Purpose. To develop investment plan of investors’ profiles taking into account their risk preferences using data
analysis of roboadviser service.
Materials and Methods. SWOTanalysis of roboadvice (RA) services and comparative characteristics of
roboadvisers explain advantage of RA services. Microservice for calculating stable consumption, finance con
sulting model of roboadvisor to ensure a constant level of consumption for the client are developed using the
following technologies: Python 3.6, Django 2.0, Django Rest framework, AngularJs, HTML5, CSS 3, Bootstrap.
Results. We considered consumptionsaving ratio in economics, emerging trends of roboadvice (RA) services
for making investment decisions. A mathematical model of roboadvisor in longrun period was developed and the
support of investment decision making was described using microservice of roboadvisor.
Conclusions. The development RA is intended primarily for private persons (investors) who invest in long
term financial instruments in order to provide them with a permanent passive income based on their chosen sa
vings period and the moment of retirement.
K e y w o r d s : roboadvisor, data analysis, long life decision making, annuity.
the problem of the optimal balance between consumption and savings, transformed into
investments, is one of the most important issues on all levels of economic system. this is
explained by the fact that the equivalence between consumer and savings flows provides
for internal and external equilibrium in the economics, and, therefore, a balanced eco-
nomic growth and increase of economic and social welfare. as a result, the scholarly ap-
ISSN 1815-2066. Nauka innov. 2020. 16(2)
kobets, V., yatsenko, V., Mazur, a., and Zubrii, M.
88 ISSN 1815-2066. Nauka innov. 2020. 16 (2)
paratus of the subject is characterized by the
variability of approaches and views of top econo-
mists, which sometimes supplement, and often
directly contradict each other. however, the bott-
leneck of the scientific research is the analysis of
consumer spending and saving patterns through
adaptive or rational expectations theories. usage
of these theories is complicated in modern econo-
mic conditions, characterized by high levels of fu-
ture uncertainty, volatility of the main econo mic
indicators, variability of market condition and
limited ability of a person to process modern data
independently, etc. in this case, the ability of eco-
nomic actors to make rational economically justi-
fied solutions dramatically reduces which deter-
mines the necessity to use automated systems for
making investment decisions. the most popular
ones among such automated systems are robo-
advice services which have the mathematical al-
gorithm based on the main principles of consump-
tion-savings theories.
the purpose of this paper is to develop invest-
ment plan of investors’ profiles taking into account
their risk preferences using data analysis of robo-
adviser service.
RoBo-advice SeRviceS
and itS eMeRGinG tRendS
the concept of “robo-advice” means the use of auto -
mation and digital techniques in order to build
and manage portfolios of exchange-traded funds
(EtFs) and other instruments for investors [1].
on the other hand, it can be described as the sys-
tem to make investment decision exclude human
psychology or emotion [2] by limitation of the trap
of repeating unfavourable previous decisions [3];
an artificial intelligence system that makes deci-
sions based on algorithms by collecting vast quan-
tities of big data [4]. the main principle of robo-
advice services means completing a simple profile
and risk tolerance questionnaire online and re-
ceive a recommended portfolio, composed mostly
of low-cost exchange-traded funds (EtFs).
robo-adviser represents an online service that
helps the client to form an investment portfolio
and subsequently manage it (make adjustments,
advice, etc.). the consulting robot is able to ana-
lyze the user's needs and his/her risk attitude,
make an investment model for him and gradually
implement it by buying and selling securities in
the stock market and other financial instruments.
Table 1. SWOT-Analysis of Robo-Advice
Services Analysis
strengths
increased productivity
low commission and fees
high accessibility by customers
streamlining the account opening process, ability to
transfer assets
no requirement for a deep financial background
increased transparency
Diversification of portfolios by using EtFs
it is the ideal model for clients with simple needs
Monitoring, rebalancing and reporting on portfolios
appealing to non-traditional clients (young clients)
offers easy-to-use tools simplifying the client experience
opportunities
adding new capabilities;
attracting assets that are not currently in-house at wealth
management firms;
receiving synergy and added value in case of cooperation
with financial advisers;
Expanding interest in passive investing;
assimilating multiple goals (college savings, retirement,
estate planning and the need for health care and/or long
term care coverage);
including extra assets (equities, fixed income, hedge funds
and real estate);
helping clients to understand their portfolios;
Weaknesses
it does not meet all needs of investors
it does not suit every investor
it uses simple surveys to profile clients and to assess their
needs
it proposes fairly basic capabilities
it has minimal ability to explain complex topics, and no
ability at all to follow up with questions and make re com-
mendations based on the answers
threats
Failure of trust to automation and digital techniques
necessity of face-to-face interaction between clients and
advisors
Data analysis of personalized investment Decision Making using robo-advisers
ISSN 1815-2066. Nauka innov. 2020. 16 (2) 89
thus, the robo-adviser actually performs the func-
tions of the "portfolio manager".
the income depends on the amount invested
and the degree of investment risk. it is impossible
to guarantee a positive and stable profitability
when trading financial instruments on the stock
exchange. this is a feature of the industry itself,
not just this service.
the scale of asset management with robo-advi-
sor continues to be quite insignificant. according
to research of Jae yeon park and et al. it is two
hundred trillion won, while it is expected more
than two thousand trillion won in 2020 [2]. accor-
ding to Deloitte prediction robo-advisory ser vices
will manage with assets between usD 2.2 tril lion
and usD 3.7 trillion in 2020. By the year 2025
is expected to rise to over usD 16.0 trillion as-
sets under management [5]. swot-analysis of ra is
de monstrated in table 1.
uptrend of robo-advisor development can be
explained due to its main parameters: strike pri-
ce, variety and complexity of investment tools
and trust. We assume that new capabilities of ro-
bo-advisors and trust will be developed more ac-
tively and rapid then the price and commission
of robo-advice services which means exponential
Table 2. Comparative Characteristics of Robo-Advisers
Day robo-advisor Features and positioning Marketing stock
Minimal
amount
Commission
Best in total
indicators with
the function of
minimizing taxes
Wealthfront optimization and inde-
xation of taxes for accounts
over usD 100 000
up to usD 10 000
no commission for
annual maintenance
usD 500 0.25% per year
Best in total
indicators
Betterment With the help of a user-
friendly interface, it is
possible to form a
portfolio efficiently and
cheaply
First 6 months free
of charge
usD 0 0.15—0.35% per
year
With minimal
commissions
WiseBanyan additional services paid none usD 10 0
With minimal
commissions
Charles
schwab
recognized leader in the
sphere of asset
management
— usD 5 000 0
specializing in
retirement plan
401k
Futureadvisor operates 401k accounts
serviced by the
investment company
Fidelity and tD
ameritrade for free
First 3 months free
of charge
usD 10 000 0.5% per year
specializing in
retirement plan
401k
Blooom Manages 401k accounts
for a fixed fee
none usD 0 usD 5 to usD 99
per month
With the services
of a financial
adviser
Vanguard services of the adviser
(manager)
— usD 50 000 0.30% per year
With the services
of a financial
adviser and
function of
minimizing taxes
personal
Capital
Financial Consultant
(Manager) + tax
optimization for accounts
over usD 100 000
— usD 25 000 0.49–0.89% per year
kobets, V., yatsenko, V., Mazur, a., and Zubrii, M.
90 ISSN 1815-2066. Nauka innov. 2020. 16 (2)
growth and expansion robo-advisors among dif-
ferent types of investors. nevertheless, now trust
is one of the barriers of significant robo-advice
ser vices growth. accenture research indicates that
77% of wealth management clients trust their fi-
nancial advisors and want to work with them to
grow and manage their wealth. Furthermore, 81%
say that face-to-face interaction is important—
the highest figure of all channels [1]. Customers
prefer hybrid solutions, allowing them to search
for information and compare available products
online, but still request human advisory before
committing to an investment [6].
the first such services for retail customers ap-
peared in the us, in 2008. according to the re-
searchhQ news specializing in research and ra -
tings, the leaders of this market are american in de-
pendent robot-advisers Wealthfront (usD 3 bil lion
under management) and Betterment (usD 4.2 bil-
lion, 150 thousand customers). in recent years, the
number of such online advisers has increased no-
ticeably. there are more than 200 firms offering
services of robo-acquiring around the world. no-
wadays, they manage about usD 300 billion and,
according to forecasts, this graph will grow.
Comparative characteristics of robo-advisers are
given in table 2 and collected from internet por tals
that analyze foreign brokerage services. the table in
the left column, has the criteria, for which two lea-
der companies are allotted, highlighted.
1. the commissions are charged for the use of
robot-advisers, but the costs for the purchase of
assets are not taken. Most of robot-advisers use
EtFs (Exchange traded Funds), and managing
companies that issue EtFs, charge 0.1 to 0.5%
per annum on funds, depending on the composi-
tion of assets. therefore, it is necessary to add the
specified commissions to these prices.
2. optimization of taxes: the company studies
the client's tax profile for the previous period in
order to find inefficiently paid deductions (pen-
sion, insurance, banks, medicine, stock market),
return them and reinvest to the client's account.
3. retirement plans 401k is a popular tool in
the us for saving. Money is transferred to these
accounts by employers, the assets are not in pen-
sion funds, but on dedicated accounts in invest-
ment companies, taxes are not withheld from in-
vestments. Many people use several options at
once, and often the calculation is based on house-
hold income.
WHo WiLL BeneFit?
in well-developed countries, due to significant
amount of savings and mature financial markets,
the great majority of people have an investment
interest, however individual investors usually
make investment decisions based on their limited
professional knowledge and asymmetric informa-
tion while institutional investors have stronger
professional background and informational sup-
port. in this case, robo-advisor is the best tool for
private investors the main aim of which is to save
the current value of their wealth.
nevertheless, Jae yeon park and et al. pointed
out that institutional investors are always ex-
posed to risks caused by diverse unpredictable
variables in the financial market [2] which makes
robo-advice service suitable also for professional
investors with huge amount of investments. in
this instance, by applying the pareto principle,
we can assume that robo-advisors can be used in
fixed proportions for different types of investors
(Fig. 1).
For example, 80% of investments assets of pri-
vate investors can be under robo-advice services
due to its low commission and using quite simple
and clear financial instruments. this stra tegy ac-
cording to pareto-law will gain 20% of all invest-
ment profit, which enables meeting main aim of
private investor — save their capital. the rest 20%
of assets can be invested in risky investments with
the help of experienced financial consultants,
which can receive 80% of investment profit.
Vice versa, for institutional investors only 20%
assets can be under robo-advice services because
of its immature and narrow range of financial and
investment instruments since now it does not
meet the needs of investors with even moderately
complex financial lives [1]. at the same time, it
Data analysis of personalized investment Decision Making using robo-advisers
ISSN 1815-2066. Nauka innov. 2020. 16 (2) 91
can gain 80% of all investment profit which al-
lows institutional investors to meet all clients’
needs in long-run period and keep profitability.
the rest amount of assets can be invested in com-
pound instruments based on professional back-
ground and classified market information.
such kind of combination is the most effective
to our mind, since complete substitution of hu-
man interaction in investing process is impos-
sible. Firstly, robo-advisor capabilities do not
co ver the whole range of all latest financial and
investment instruments, secondly, automated al-
gorithms of a robo-advisor is based on self-repor-
ting processes which can lay down a wrong prin-
ciple of the robo-advisor working due to incorrect
investor self-assessment.
Privat
investor
Aim
Institutional
investor
80%
Robo-advice service
20%
Robo-advice
service
20%
Financial
consultants
80%
Financial consultants
20% of profit
80% of profit
Multipli-
cation
of capital
Saving
of capital
Simple tools + asymmetric
information + low commission
Complex investment tools + professional
background + access to the market infomation
Fig. 1. the possibility of using robo-advisors by different types of investors
Purpose of investing
Type of investor
Begining
Investment decision
Various investment tools
Various investment tools
Welth Save welth
Robo-advisory
The end
Institutional investor Private investor
Fig. 2. robo-advice services
kobets, V., yatsenko, V., Mazur, a., and Zubrii, M.
92 ISSN 1815-2066. Nauka innov. 2020. 16 (2)
on the other hand, we can mentioned that due
to its adaptiveness robo-advice services can be
useful for both aggressive and conservative types
of investors for which there are two types of rob-
advisories: active (the investor only receives re-
balancing suggestions and decides self-directly
about actual execution) and passive (rebalancing
is fully quantitative) [6]. the latest generation
(robo-advisor 4.0) shifts between different asset
classes based on changing market conditions and
individual investment needs such as profit, risk
appetite, and liquidity aspects, monitor and ad-
just single client portfolios in real time to keep on
track with their selected investment strategy [5].
analyzing short and long-term oriented in-
vestments, it can be easily assumed that the robo-
advice capability is indeed the better advisor in
the long run since the its mathematical algo -
rithm propose adequate investment opportuni-
ties based on stable economic trends, excluding
the short-term speculative and human being mar-
ket precedents.
Finally, we can sum up that in the current stage
of robo-advisors development this investment
Table 3. Description of the Features for Clients
of Robo-Adviser Services
Feature type Domain
risk profile ordinal [very low; low; normal; high; very
high]
investment
goals
ordinal [very low; low; normal; high; very
high]
sex Binary [male, female]
age integer [17…90]
Fig. 3. Workflow of decision making for robo-adviser services
Result
Final Portfolio Ranked Portfolios
Neighborhood
New Client
tool is more suitable for private investor who is
oriented for long-run investment horizon (Fig. 2).
that is why the robot-adviser developed by us
is designed primarily for private persons (inves-
tors) who invest for a long time in financial inst-
ruments in order to secure a permanent passive
income at the planned retirement age with the
help of a robot consultant.
cLUSteRinG aLGoRitHM FoR inveStoRS
let’s consider workflow of robo-adviser services
(Fig. 3). Each user is represented as a vector of
four features: risk profile, extracted via standard
MiFiD questionnaire, investment goals, sex and
age. the first two features are described on five-
point ordinal scale, from very low to very high,
whi le sex is represented in a binary fashion (ma-
le = 0, female = 1), age is characterized as a nu-
merical variable (table 3).
to distinguish different types of investors (new
clients) we can use cluster analysis which help us
to reveal main types of most similar investors
(neighborhood). after that we can construct cor-
respondence between specific users and invest-
ment portfolios. then we choose best final port-
folio which corresponds described features of in-
vestor using profitability index (investment goal)
and risk attitude (risk profile). ra’s service auto-
matically invests in preferred financial instru-
ments to achieve investment goal of client for
given investment horizon. Finally client can com-
pare planned goal and achieved result, and revise
him or her decision about own investment prefe-
rences (profitability and risk).
k-means clustering algorithm for investors con-
sists of following steps:
1) Choosing number of clusters Ek for given da-
ta set of investors;
2) selection of initial features for reference da-
ta which correspond to number of clusters;
3) Calculation distance between features of in-
vestors and reference data using following for-
mula (1):
d(Ai;Ek) = (xj —ej)2, (1)
j = 1
m
∑
Data analysis of personalized investment Decision Making using robo-advisers
ISSN 1815-2066. Nauka innov. 2020. 16 (2) 93
where i — index of investor; xj — feature j of in-
vestor i; ej — initial features for reference data;
n — number of investors, m — number of charac-
teristics of each investor;
4) Changing coordinates for reference data Ek
according to minimal distance between given fea-
tures of investors and reference data;
5) Defining final values of reference data Ek;
6) preparing the classification of objects.
Main types of investors include risk-averse in-
vestors (2), risk seeking investors (3) and hybrid
type of investors (4) which combines previous
ones. these types of investors are correspon-
dingly described by following conditions [24, 25]:
using cluster of clients and types of investors
we map set of users in set of investors’ types as
described in following section.
a wide series of experiments has been carried
out to validate the types of investors for ra ser-
vice using dataset of 100 randomly chosen real
(anonymous) users, who choose portfolios be-
tween June 2011 and June 2013. this dataset is
available for download by objectway Financial
software (http://www.di.uniba.it/swap/finan-
cialrs_data_uniba.zip).
Data set of features for users which invest in
portfolio of finance instrument is described on
Fig. 4, a. in this case we quantified our ordinal data:
risk profile = [very low; low; normal; high; very
high] = [1, 2, 3, 4, 5];
investment goals = [very low; low; normal;
high; very high] = [1, 2, 3, 4, 5],
sex = [male, female] = [0, 1].
average values of indicated features are follow-
ing: risk profile = 3.16 (closely to normal), in-
vestment goals = 3.08 (almost normal), male =
56% (female = 44%) (Fig. 4, a), age = 66.8 year
(Fig. 4, b).
k-means clustering algorithm for investors re-
veal 2 main clusters for our dataset:
1) active working age and risk-seeking person
(man as a rule) who require more than average
income (52%);
Fig. 4. Characteristics of investment portfolio: a — Data set of
features for users of ra; b — Distribution of investors' age
a
b
30
Number of investors
40
70
80
90
Age
50
60
1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96
i = 1
n
∑
j = 1
n
∑rp = xi · xj · vij → min
i = 1
n
∑ xi · di = mp , (2)
i = 1
n
∑ xi = 1, xi ≥ 0.
{
i = 1
n
∑
j = 1
n
∑rp = xi · xj · vij → min
i = 1
n
∑ xi · di = mp , (3)
i = 1
n
∑ xi = 1, xi ≥ 0.
{
(4)
i = 1
n
∑ xi = 1,
xi ≥ 0.
{ xi
2 · vi
2 + xi xj vij
i = 1
n
∑ xi · di
→ min,f (xi) =
∑ i= 1
n ∑ i= 1
n ∑ i= 1
n√
kobets, V., yatsenko, V., Mazur, a., and Zubrii, M.
94 ISSN 1815-2066. Nauka innov. 2020. 16 (2)
–0.4
Time series (day 23.06.18–0.5.08.17)
BTC-USD
ETH-USD
LTC-USD
NEO-USD
BCH-USD
R
at
e
of
c
ry
pt
oc
ur
re
nc
ie
s
–0.2
0.4
0.6
0.8
0
0.2
1 12 23 34 45 56 67 78 89 100 111 122 133144 155166 177 188 199 210 221 232 243254 265276 287 298 309 320
2) retirement pensioner and risk-neutral per-
son (often it is a man) who need income less than
average (48%).
using open data about rate of financial inst-
ruments (Fig. 5), such as cryptocurrencies
(https:// finance.yahoo.com/cryptocurrencies)
and Marko witz model we developed 3 main types
of investors:
1st type of investor: daily profitability of the
risk-averse investor is 0.657% daily or 24.5%
annually with 0.44% risk;
2nd type of investor: for the risk-seeking inves-
tor, the yield will be higher — 0.87% daily or
31.6% annually with 1.06% risk;
3rd type of investor: for hybrid type of inves-
tor the yield will be 0.45% daily or 16.5% an-
nually with 0.28% risk.
thus active working age and risk-seeking per-
son is mostly corresponding to 2nd type of inves-
tor (profitability index = 31.6% for robo-advisor
service). at the same time retirement pensioner
and risk-neutral person is associated with 3rd type
of investor (profitability index = 16.5% for robo-
advisor service).
MatHeMaticaL ModeL
oF RoBo-adviSoR
in LonG RUn PeRiod
let’s consider the work of robo-advisor service
based on the following example. task of the ser-
vice is to determine the conditions under which
an investor can carry out regular consumption
both before and after retirement using a personal
savings fund. to achieve this, the client of this
service should answer the following questions:
1. What is the average annual income of the
client?
2. From what age does the client plan to start
a personal savings program?
3. at what age does the client plan to retire?
4. up to what age does the client plan to use
his/hers personal savings fund?
5. Which average level of ‘risk-profit’ (‘h-i’) for
financial instruments is preferable for the client?
Fig. 5. rate of 5 main cryptocurrencies for users of ra (23.06.18-05.08.17)
Data analysis of personalized investment Decision Making using robo-advisers
ISSN 1815-2066. Nauka innov. 2020. 16 (2) 95
6. What is the acceptable level of annual con-
sumption C* for the client?
the task assignment of the it service is to
maintain a constant level of client’s consumption
during life-long period through automated analy-
sis of how much he/she has to consume and save
each year. Example of necessary indicators for
data analysis is presented in table 4. Calculation
of distributed income Y for consumption C and
savings S:
C = Y — S. (5)
the task assignment of robo-advisor consists
of two parts:
1. to define the annual amount of savings using
the future value of annuity FVA:
FVA= S × , (6)
where i is the desired annual real interest rate on
savings, N1 = t2 — t1 is the accumulation period of
a personal savings fund, S is the annual amount
saved. after N1 years the personal savings of the
client will equal FVA.
2. to calculate the distribution of the savings
fund on constant consumption after retirement
using the present value of annuity:
PVA = C × , (7)
Where і is scheduled annual real return on sa-
vings, N2 = t3 — t2 is utilization period of a per-
sonal savings fund, C is annual constant consump-
tion of the client.
thus the future and present values of annuities
should be equal: FVA = PVA (future and present
values of annuities), that is, taking into account
(5) we obtain:
(8)
after substitution of the values from Тable 4 in
formula (8) we get: C = Eur 182411.7 is the an-
nual constant level of consumption, which is not
less than the desired expenditure level on the an-
nual consumption of the client C* = Eur 18000.
at the same time, the annual level of client’s sav-
ings should be Eur 17588.3. if i << 100%, then
we can rewrite (8) as following and
savings fund is .
the formation of a personal consumption fund
in graphic form is presented in Fig. 6, a. Distribu-
Table 4. Indicators of Robo-Advisor Service
indicators Value of indicators
the age of the client from which the
personal savings program begins
t1 = 35 years
retirement age of the client t2 = 65 years
age up to which the client will use
his savings fund
t3 = 80 years
annual income of the client Y = Eur 200000
the desired annual real interest for
risk-free investments
i = 7%, h = 0
acceptable level of annual con sump-
tion of the client
Y = Eur 180000
Fig. 6. Future and present values of annuity (FVA and PVA): a — Formation of a personal consumption fund FVA; b —
Distribution of personal savings fund PVA
17588.3
1
1
2
2
0
0
29
3
30
15
...
...
125127.5
116941.6
18819.5
FVA-1661406
FVA-1661406
17588.3 17588.3 17588.3 17588.3
182411.7 182411.7 182411.7 182411.7 182411.7
170478.2
159325.5
148902.3
66114.4
(1 + i) N1 — 1
i
(1 + i) —N2
i
Y ((1 + i) N1 — 1)
(1 + i) N1 — (1 + i) —N2
C = .
Y · N1
N1 · N2
C =
Y · N2
N1 · N2
C =
kobets, V., yatsenko, V., Mazur, a., and Zubrii, M.
96 ISSN 1815-2066. Nauka innov. 2020. 16 (2)
tion of personal savings funds to a constant con-
sumption level of Eur 182411.7 is presented in
Fig. 6, b.
SUPPoRt oF inveStMent deciSion
MaKinG in LonG-RUn PeRiod via
SoFtWaRe ModULe oF RoBo-adviSoR
to implement the practical part we developed a
program module as web cloud service of robo-
advisor using python Django technology and temp-
late of user interface admin ltE. in its final form
the robo-advisor will have the architecture de-
picted in Fig. 7.
the overview of the practical module is the fol-
lowing: the application is presented in following
figures. Firstly, user needs to pass the registration
or if he has already registered — log in to the ser-
vice (Fig. 8).
then, the user has to pass two questionnaires
on the website.
on first questionnaire page, the user has to en-
ter firstly the input data required for further cal-
culations, such as:
risk profile;
investment goals;
sex;
age.
the system will use these data for cluster anal-
ysis to know user’s risk level (Fig. 9).
then user has to pass the second questionnaire
and answer next questions:
From what age do you want to start your per-
sonal savings program?
at what age do you plan to retire?
to what age do you plan to use your personal
savings program?
What is your average annual revenue?
Fig. 7. robo-advisor architecture
users
users data
Data source
Clean up
module
investment
plan module
security
module
Calculation
module
Frontend
parset data
parser
Fig. 8. robo-advisor creation process
registration page log in page
Data analysis of personalized investment Decision Making using robo-advisers
ISSN 1815-2066. Nauka innov. 2020. 16 (2) 97
What average profitability of financial instru-
ments do you expect?
What risk level of financial instruments is ac-
ceptable for you to place your savings to?
What is acceptable level of spending on your
annual consumption?
the system includes methods for calculating S
(payment of customer), C (permanent consump-
tion of client), FVA (Future Value of an annui-
ty), PVA (present Value of an annuity), provided
in the application model, which uses input data,
provided by the user (Fig. 10). after calculations,
the system displays the calculated savings S and
consumption C to the user during long-life peri-
od, which is based on the input data.
in the Fig. 10 are displayed results based on re-
sults of cluster analysis for representative clients:
interest per year = 31.6% and risk per month =
= 32.3%. user can modify the questionnaire data
not more often, than once a week and system will
automatically re-calculate the result based on new
data. if payment of customer will be changed —
system will display it in a graph as a new break-
point.
system will automatically send emails with no-
tifications to customers if profitability of finan-
Fig. 9. First questionnaire about investor profile
Fig. 10. robo-advisor: input and output
Data input output data
kobets, V., yatsenko, V., Mazur, a., and Zubrii, M.
98 ISSN 1815-2066. Nauka innov. 2020. 16 (2)
cial instruments will be changed and these notifi-
cations will be displayed in the system. user can
choose new investment plan and system will show
potential profitability changes. For high-level
risk users system will send it every day, for hy-
brid-level risk users — twice a week, for low-level
risks users — once a week.
concLUSionS
in our opinion, savings are remaining balance of
disposable income directed to maintaining or im-
proving the standard of living in the future. this
fact describes an additional argument in favor of
using an automated financial consultant who will
assess the necessity of an immediate change in
the average propensity to consume, depending on
the model parameters changes. the concept of “ro-
bo-advice” means using of automation and digital
techniques to build and manage portfolios of ex-
change-traded funds and other financial instru-
ments for investors. robo-adviser represents an
online service that helps the client to form an in-
vestment portfolio and subsequently manage it.
the robot-adviser developed by us is designed
primarily for private persons (investors) who in-
vest for a long time in financial instruments in
order to secure a permanent passive income at the
planned retirement age with the help of a robot
consultant. the task assignment of our it service is
to maintain a constant level of client’s consump-
tion during life-long period through automated
analysis of how much he or she has to consume
and save each year. to define the annual amount of
savings to guarantee constant level of consump-
tion for each private person during long life pe-
riod we use present and future value of annuity.
in the future it is planned to expand the system
in following directions: develop a web-based ver-
sion of the application, add integration with ex-
ternal services to find financial instruments for
investment, add administration system and im-
prove the application interface.
rEFErEnCEs
1. the rise of robo-advice. Changing the Concept of Wealth Management. url: https://www.accenture.com/_acnme-
dia/pDF-2/accenture-Wealth-Management-rise-of-robo-advice.pdf (last accessed: 16.12.2018).
2. park, J. y., ryu, J. p., shin, h. J. (2016). robo advisors for portfolio Management. Adanced Science and Technology
Letters, 141 (Green and smart technology ii), 104—108. doi: dx.doi.org/10.14257/astl.2016.141.21
3. alós-Ferrer, C., hügelschäfer, s., li, J. (2016). inertia and Decision Making. Front. Psychol., 7, 169. doi: 10.3389/
fpsyg.2016.00169
4. park, J. h., ryu, J. p., shin, h. J. (2016). predicting kospi stock index using Machine learning algorithms with
technical indicators. Journal of Information Technology and Architecture, 13, 331—340.
5. the expansion of robo-advisory in Wealth Management — Deloitte. url: https://www2.deloitte.com/content/dam/
Deloitte/de/Documents/financial-services/Deloitte-robo-safe.pdf (last accessed: 16.12.2018).
6. Jung, D., Dorner, V., Glaser, F., Morana, s. (2018). robo-advisory — Digitalization and automation of Financial advi-
sory. Business & Information Systems Engineering, 60(1), 81—86. doi: 10.1007/s12599-018-0521-9
7. kohavi, r., provost, F. (1998). Glossary of terms. Machine learning — special issue on applications of Machine learning
and the knowledge Discovery process. Machine Learning, 30, 271—274.
8. the implications of machine learning in finance. url: https://www.bloomberg.com/professional/blog/implications-
machine-learning-finance/ (last accessed: 16.12.2018).
9. lam, J. W. (2016). robo-advisers: a portfolio Management perspective. Senior Thesis, Yale College. url: http://eco-
nomics.yale.edu/sites/default/files/files/undergraduate/nominated%20senior%20Essays/2015-16/Jonathan_lam_
senior%20Essay%20revised.pdf
10. Faggella, D. Machine learning in Finance — present and Future applications. url: https://www.techemergence.com/
machine-learning-in-finance/ (last accessed: 16.12.2018).
11. kashner, E. Ghosts in the robo advisor Machine. url: http://www.etf.com/sections/blog/22973-ghosts-in-the-robo-
advisor-machine.html (last accessed: 16.12.2018).
12. Markowitz, h. M. (1952). portfolio selection. The Journal of Finance. 7(1), 77—91. doi:10.2307/2975974. Jstor
2975974
Data analysis of personalized investment Decision Making using robo-advisers
ISSN 1815-2066. Nauka innov. 2020. 16 (2) 99
13. Fisher, i. (1977). The Theory of interest. philadelphia: porcupine press.
14. Black, F., litterman, r. (1992). Global portfolio optimization. Financial Analysts Journal, 48(5), 28—43. doi:10.2469/faj.
v48.n5.28
15. Baker, t., Dellaert, B. (2018). regulating robo advice across the Financial services industry. Faculty scholarship at
penn law. 1740. url: https://scholarship.law.upenn.edu/faculty_scholarship/1740 (last accessed: 16.12.2018)
16. Betterment review. url: https://www.nerdwallet.com/blog/investing/betterment-review/ (last accessed: 16.12.2018).
17. Future advisor review. url: https://www.nerdwallet.com/blog/investing/futureadvisor-review/ (last accessed: 16.12.2018).
18. thangavelu, p. Motif investing Broker review: Easy thematic investing. url: https://www.investopedia.com/arti-
cles/active-trading/030415/motif-investing-broker- review-easy-thematic-investing.asp (last accessed: 16.12.2018).
19. Motif investment review. url: https://www.nerdwallet.com/blog/investing/motif-investing-review-1/ (last acces-
sed: 16.12.2018).
20. Fein, M. l. robo-advisors: a Closer look. url: http://dx.doi.org/10.2139/ssrn.2658701 (last accessed: 16.12.2018).
21. robo advising — kpMG. url: https://home.kpmg.com/content/dam/ kpmg/pdf/2016/07/robo-advising-Catching-
up-and-Getting-ahead.pdf (last accessed: 16.12.2018).
22. kobets, V., yatsenko, V. (2016). adjusting business processes by the means of an autoregressive model using BpMn 2.0.
CEUR Workshop Proceedings, 1614, 518—533. url: CEur-Ws.org/Vol-1614/iCtEri-2016-CEur-Ws-Volume.pdf
(last accessed: 16.12.2018).
23. kobets, V., poltoratskiy, M. (2016). using an Evolutionary algorithm to improve investment strategies for industries
in an Economic system. CEUR Workshop Proceedings, 1614, 485—501. url: CEur-Ws.org/Vol-1614/iCtEri-2016-
CEur-Ws-Volume.pdf (last accessed: 16.12.2018).
24. snihovyi, o., ivanov, o., kobets, V. (2018). implementation of robo-advisors using neural networks for Different risk
attitude investment Decisions. 9th International conference on intelligent systems, (25—27 september 2018, Funchal-
Madeira, portugal), Funchal-Madeira, 2018, 332—336. doi:10.1109/is.2018.8710559 (last accessed: 16.12.2018).
25. kobets, V., yatsenko, V., Mazur, a., Zubrii, M. (2018). Data analysis of private investment Decision Making using tools
of robo-advisers in long-run period. CEUR Workshop Proceedings, 2104, 144—159. url: CEur-Ws.org/Vol-2104/
(last accessed: 16.12.2018).
Стаття надійшла до редакції / Received 04.01.19
Статтю прорецензовано / Revised 10.04.19
Статтю підписано до друку / Accepted 07.05.19
В.М. Кобець1, В.О. Яценко2, А.Ю. Мазур1, М.І. Зубрій1
1 херсонський державний університет,
вул. Університетська, 27, херсон, 73003, Україна,
+380 552 32 6782, kipiekedu@gmail.com
2 київський національний університет імені Тараса шевченка,
вул. Васильківська, 90-а, київ, 03022, Україна,
+380 44 521 3578, decanat_econom@univ.kiev.ua
аНаЛІз ДаНих ПерсоНаЛІзоВаНоГо ПрийНяТТя
ІНВесТицІйНих рІшеНь з ВикорисТаННяМ
аВТоМаТизоВаНих ФІНаНсоВих коНсУЛьТаНТІВ
Вступ. На сьогодні проблема пошуку оптимального балансу між споживанням та заощадженнями, перетвореними в
інвестиції, вирішується за допомогою автоматизованих систем прийняття інвестиційних рішень, прикладом яких є
послуги автоматизованих фінансових консультантів або робо-консультантів (robo-advisers), які працюють на базі
математичного алгоритму, що ґрунтується на теорії споживання та заощадження.
Проблематика. завданням розробленого ІТ сервісу є підтримка постійного рівня споживання клієнта протягом
всього періоду життя шляхом автоматичного аналізу того, скільки він/вона має споживати та заощаджувати щороку.
результати пропозицій щодо співвідношення споживання та заощадження можуть змінюватися при зміні початкових
фінансових даних.
Мета. розробити інвестиційний план для профайлів інвесторів з урахуванням їх схильності до ризику за допомогою
аналізу даних сервісом автоматизованих фінансових консультантів (robo-advisers).
|