Forecasting of UPI Payment Services Demand in India Using Machine Learning Techniques
Dr. Vengalarao Pachava
Assistant Professor,
School of Business Management,
NMIMS University Jadcherla Campus
Email: vengalarao.pachava@nmims.edu
Dr. Siva Krishna Golla
Assistant Professor,
School of Management Studies,
National Forensic Sciences University,
Gandhinagar, Gujarat.
Email: sivakrishna.golla@nfsu.ac.in
Prof. Kavitha A Karkera
Associate Professor,
School of Management MBA,
Nagarjuna Degree College, Bengaluru,
Email: kavithakarkera@nagarjunadegreecollege.in
Dr. Abhilash Ponnam
Associate Professor,
School of Business Management,
NMIMS University Jadcherla Campus
Email: abhilashponnam@gmail.com
Abstract
Unified Payments Interface (UPI) is a technology that integrates several bank accounts into a single mobile app (of any participating bank), combining many banking services, smooth money transfer, and merchant payments under one umbrella. Because of demonetization and the Covid-19 outbreak, UPI payments have changed and increased significantly, becoming a key element of the Indian payment system. In 2022 financial year, it processed over 45 billion transactions worth over Rs.83 lakh crore. The Auto Regressive Integrated Moving Average (ARIMA) approach is well-known for predicting time series data while taking into account the non-linearity of the data. Since 2016, both the total number of transactions and the total value of transactions in the UPI system have been steadily growing. Using the ARIMA time series approach, this research forecasted the future one-year value and volume of UPI payments. The findings of this study indicate that there will a rapid growth in the use of UPI for digital transactions, large number of payments settled through the UPI platform, the trend would continue for the next calendar year.
Keywords: Unified Payments Interface, ARIMA, Banking and Time Series
Introduction
The Indian banking sector constantly evolving, and a major impetus came from the nationalization of commercial banks with social objectives; consequently, it has been witnessing a wide range of policy-induced reforms and structural changes since the early 1990s (Gupta, 2021). The nationalization of commercial banks with social objectives was a major impetus for the evolution of the Indian banking sector. The banking industry has undergone a significant shift as a result of the tremendous development made in the area of information technology(Bhuyan et al., 2021). Telecommunication and electronic data processing advancements have pushed these transformations even farther forward. The banking and financial industries throughout the world have been completely transformed as a result of automation(Usmonova Durdona Shukhratovna, 2021). Click-and-order banking channels such as online banking, automated teller machines, tele-banking, and mobile banking are becoming more popular alternatives to traditional brick-and-mortar bank branches. A few clicks are all it takes for customers to access their accounts, check their statements and move payments(Kaur et al., 2021). Due to technological advancement, availability of ICT infrastructure, improve of basic education and a sudden effect of covid 19 pandemic caused for rapid usage of e-banking features in India(Ahmed & Sur, 2021). According to a study conducted on mobile payment services such as mobile wallets and mobile banking during India's demonetization, the majority of urban youth have adopted mobile payment methods (Chopra, 2017). Immediate Payment Service (IMPS) transaction value grew by 196.7 percent year-on-year in January 2017. In December 2016, NACH (National Automated Clearing House) fund clearing platform set up by NPCI has grew 116.7%. (Sinha et al., 2019). Mobile payment penetration increased in India after the demonetization of high-value currency notes; however, usage and retention remain low; the primary reason for this is the privacy concerns associated with mobile banking. All this analysis reveals that demonetization boosted digital payments immediately, but not long-term (Chakrabarty et al., 2020).
Electronic banking, sometimes known as e-banking, is a product of globalisation, increased competition, and the explosive rise of information technology systems. It has evolved into a self-service delivery channel that enables banks to give information and services to their clients in a more convenient manner using a variety of technological services like as the Internet and mobile phones (Kurnia et al., 2010). Electronic funds transfer for retail purchases, automatic teller machines (ATMs), and automatic payroll deposits and bill payments are some of the features of e-banking. RBI initiative in India has achieved steady growth in E–Payments during the last ten years. Many e-payments systems have been developed to date in order to digitalize the current banking system. One of them is UPI (Unified Payment Interface) introduced by National Payments Corporation of India (NPCI) in the year of 2016, with the goal of simplifying, streamlining, and enhancing the security of the electronic banking sector. NPCI bringing change in all electronic payments made in India, its guidance and support by the Reserve Bank of India (RBI) and the Indian Bank Association (IBA) provided a Unified Payments Interface (UPI) system by which 24*7 payments are made easy, real-time, and frictionless (Rai et al., 2017.).
Unified Payments Interface (UPI) has secured immense recognition, a convenient device like smart phone used for diversified financial transactions. Direct transactions of payment and receipts rightly made using a virtual payment address on the UPI platform. For using a UPI interface, it is mandatory to have a Bank account and link it with the UPI application. Google Pay, Phonepe, Whatsapp Pa, Amazon Pay are a few UPI applications frequently used for a wide variety of financial transactions. The rapid advancement in the use of Unified Payment Interface has opened the gateway for most of the banks to provide this service in their mobile applications (A & Bhat, 2021). Since UPI is linked to a bank account, no wallet is needed, your bank account can be used to link UPI, customers can use any bank or third-party app with UPI's compatibility across all systems. Some of the most important factors for the success of UPI are simplicity, adoption, security and cost (Shree et al., 2021).
UPI is widely accepted as a user-friendly interface for digital settlement of transactions. UPI has eliminated the multiple mediators in the settlement of financial transaction, thereby increasing its acceptance (A & Bhat, 2021). The easy access to smart phones, the convenience of identity verification online, universal access to banking and introduction of biometric sensors in phones has encouraged the positive attitude to the acceptance of UPI for promoting a less-cash culture in India (K. Devi & Devadutta Indoria, 2021). The popularity will increase for future payments, considering its convenience, safety and minimal cost. The recognition of digital payments expected to increase in line with the overall socioeconomic development of the population (Shree et al., 2021). With the forethought to encourage cashless Indian economy, UPI has helped people tremendously in transfer of funds instantly with much ease (Madwanna et al., 2021).
Time series analysis is a specialised method of examining a set of data points gathered over a specified period of time. In time series analysis, data points are recorded at regular intervals throughout a predetermined length of time rather than being recorded randomly or irregularly. However, this form of analysis is not just the collection of data over a long period of time(Katris, 2021). The increased availability of data that is not steady has made prediction analysis more difficult. Researchers and academics are still determining the optimal approach in finance and economics to overcome this obstacle. The prediction theory prompted academicians to create several prediction models such as artificial neural networks, hybrid models, and ARIMA models(Khandelwal et al., 2015). The assumption that the underlying model is linear is what makes ARMA difficult to apply to a wide variety of complex real-world time series, despite the fact that it has been quite successful. Autoregressive integrated moving average (ARIMA) is an extension of ARMA that can deal with nonstationary time series forecasting by using differencing methods to handle this challenge(Kotu & Deshpande, 2019). Differentiating approaches may be used to reduce the effects of trend components before fitting an ARIMA model when the data contain trend and heteroscedasticity. However, the vast majority of currently available ARIMA models continue to have severe shortcomings, including low accuracy, high error rate, handling non linear data etc. But still ARIMA models using vastly in many of the domains to forecast the future values. Time series forecasting has played a key role in a variety of disciplines throughout the last several decades, including stock markets (Shah et al., 2022), daily and monthly weather forecasting (Alsharif et al., 2019), crop production forecasting (Mishra et al., 2021), disease cases prediction in medical field(Liu et al., 2016), sales prediction (Prasanth Shakti et al., 2017) , production value in mechanical industry(Liang, 2009), oil production prediction (Ning et al., 2022), financial budget forecasting(Rhanoui et al., 2019) etc.
During the fiscal year (FY) 2021-22, the unified payments interface approach successfully processed payments totalling $1.09 trillion and another key milestone in the month of March 2022 it has completed 5.04 billion transactions. Because of the focus that institutions put on decision making and strategy making predicting is the key, predicting the UPI transactions value and volume of data has become a popular topic of conversation for a diverse range of individuals for digital transformation of money.
Data
UPI payments monthly transaction value (in crore) and volume (in million) data were acquired from the internet repository of the National Payments Corporation of India. The actual data is available from 2016 august to 2022 June. However, for the purpose of this study, the researchers used the data from 2018 august to 2022 June in order to omit self-transactions data from 2016 august to 2018 august.
Figure 1: UPI payment volume and value
Methodology
Box - Jenkins analysis is a systematic approach of discovering, estimating, validating, and applying integrated autoregressive, moving average (ARIMA) time series models that was used in this study. This approach is suitable for use with time series that are of medium to long duration of the data points like at least 50 observations are required.
Model Identification
In order to identify the model, the first step is to ensure that the data exhibit stationarity properties. These features ensure that the mean, variance, and autocorrelation structure do not vary over a period.
(Auto-correlation)
Decomposition technique used to identify the data stationarity properties; it consists of four subplots, which explains about the data at level, trend, seasonality and noise in the data. If trend and seasonality presented in the data that can be considered as non-stationarity. However, the autocorrelation function indicates the relationship between the present time value and its own lagged time values. If the data exhibits a trend, the autocorrelation function's lag values can rise from low to high, whereas if the data exhibits seasonality, the function's lag values can display systematic fluctuations in the graph.
ACF(n)=
Partial autocorrelation function find the correlation between current time value with its past lag values where it controls the between values effect.
PACF(n)=
There are three major components are there in ARIMA model, those are Auto Regressive (AR), Integrated (I) and Moving Average (MV). This three components are represented as p,d,q value in the model which are representing AR (p),I(d) and MA(q).
AR(p) components used to predict the current time values with its own past time lag values.
I (d) Differenced lag value of the data
First order differencing
Second order differencing
MA (q) components used to predict the current time values with its own past error lag values.
ACF graph is useful to estimate the MA components lag values and PACF component useful to estimate the AR component.
Model Development:
The ACF and PACF value models were used to identify the suitable p,d,q values for predicting future values. This research examined several p,d,q value combinations to find the best model that met all of the assumptions. This study used Akaike's information criterion (AIC), Bayesian information criterion (BIC), and maximum log likelihood values to identify the best model from the available combinations. The optimal model for predicting time series values is one that has a low AIC value and a high log likelihood value.
AIC = -2/N * LL + 2 * k/N
BIC = -2 * LL + log(N) * k
Model Validation
The assumption that there is no autocorrelation and that the distribution of the residual values follows a normal (Gaussian) distribution. In order to investigate such assumptions, this research made use of the Ljung-Box test statistic in conjunction with the autocorrelation function (ACF) of the residuals. The Kolmogorov-Smirnov normality test and the arch test used to assess normality and heteroscedasticity, respectively.
Results and Discussion
The data stationarity is a fundamental quality in the box-Jenkin approach. Initially, this study verified the data stationarity using ACF graphs of the volume and value of UPI payments. The ACF plots revealed a declining pattern of correlation between the data and its very own lag values. This indicates that the data did not adhere to stationary qualities.
Figure2: ACF plots of UPI payment volume and value
The differencing method used in order to transform this non-stationary data into stationary form; in the second differencing, the data transformed into stationary form in terms of both its volume and its value. Statistical verification of data stationarity conducted using the ADF test.
H0: Unit root presented in the data
H1: Unit root not presented in the data
The results of the ADF test statistic p value for the UPI value of 0.0001 and the UPI volume of 0.01474 shown that the null hypothesis rejected, which indicates that the data demonstrates stationarity at both differencing levels.
|
ARIMA Models |
|||||||
|
|
|
Volume |
Value |
||||
|
Characteristics |
|
(0,2,2) |
(2,2,1) |
(1,2,1) |
(0,2,2) |
(1,2,2) |
(2,2,1) |
|
AR(1) |
α1 |
|
-0.2599 |
-0.2449 |
|
0.6263 |
-0.3579 |
|
|
SEB |
|
0.1746 |
0.1594 |
|
0.1302 |
0.1688 |
|
|
Z Value |
|
-1.4884 |
-1.5370 |
|
4.8099 |
-2.1197 |
|
|
P value |
|
0.1366 |
0.1240 |
|
0.0000 |
0.0340 |
|
AR(2) |
α 2 |
|
-0.0351 |
|
|
|
-0.1439 |
|
|
SE |
|
0.1688 |
|
|
|
0.1640 |
|
|
Z Value |
|
-0.2077 |
|
|
|
-0.8773 |
|
|
p value |
|
0.8354 |
|
|
|
0.3800 |
|
MA(1) |
Ø 1 |
-1.1226 |
-0.8482 |
-0.8557 |
-1.2852 |
-1.9351 |
-0.8414 |
|
|
SE |
0.1587 |
0.0840 |
0.0737 |
0.1672 |
0.1713 |
0.0797 |
|
|
Z Value |
-7.0743 |
-10.0918 |
-11.6150 |
-7.6857 |
-11.2940 |
-10.5530 |
|
|
p value |
0.0002 |
0.0000 |
0.0000 |
0.0000 |
0.0000 |
0.0000 |
|
MA(2) |
Ø2 |
0.2423 |
|
|
0.3974 |
0.9963 |
-0.8414 |
|
|
SEB |
0.1527 |
|
|
0.1619 |
0.1758 |
0.0797 |
|
|
Z Value |
1.5863 |
|
|
2.4543 |
5.6670 |
-10.5530 |
|
|
p value |
0.1127 |
|
|
0.0141 |
0.0000 |
0.0000 |
|
AIC |
|
608.2100 |
610.3200 |
608.3600 |
1071.0000 |
1071.0400 |
1073.7200 |
|
Log likelihood |
|
-301.1000 |
-301.1600 |
-301.1800 |
-532.5000 |
-531.5200 |
-532.8600 |
|
RMSE |
|
186.8556 |
187.1650 |
187.2910 |
31827.0000 |
29823.6000 |
32138.0000 |
|
MAPE |
|
6.7453 |
6.7001 |
6.6638 |
6.6774 |
6.5519 |
6.6367 |
Table 1: ARIMA model for UPI payment value and volume
A number of models were examined with several iterations of p, q values in order to build a model that takes stationarity data into consideration. The AIC and log likelihood values were used to identify which three top models should be chosen for the final comparison. Table 1 provides an overview of the several UPI payment volume and value models that available. The ARMA (0,2) model was selected for forecasting purposes for the UPI payments volume data because, out of all the models, this model had the lowest AIC value (608.21), and all of the coefficient values were determined to be significant. Additionally, this model was the only model in which all of the coefficient values were significant. The ARMA(1,2) model was selected for the UPI payments value data because it has the AIC value (1071.04), which was the lowest of all the models, and all of its coefficient values are significant. Model residual autocorrelation tested with Ljung boxt test, where both UPI value and volume models residuals were not indicating any significant auto correlation presented in the model. Table 2 indicates the model fit values generated through Ljung box test. Test statistic p values greater than .05, where failed to reject null hypothesis can be conclude that both the models were fit.
H0: Model doesn’t show lack of fit
H1: Model shows lack of fit
Table 2: Ljung box model fitness test.
|
Model |
Test Type |
Chi-square |
Df |
p-value |
Decision |
|
Volume |
Ljung Test |
23.866 |
24 |
0.4693 |
Accepted |
|
Value |
Ljung Test |
32.277 |
24 |
0.1203 |
Accepted |
Both models projected over the next two years, and by June 2024, the forecasting graphs indicate a clear increase trend for both UPI payments volume and value. The expected value and volume of UPI payments are shown in Table 3. The volume of UPI transactions expected to expand from 5.86 billion per month to 11.41 billion by the end of June 2024. The value of UPI transactions is expected to rise from 1014384 crore to 2029901 crore rupees by the end of June 2024. This is approximately two times the value of the current month, which suggests tremendous growth in the months to come.
Table 3: UPI payments forecasted value and volume
|
|
|
Volume |
Value |
||||
|
Month |
Year |
Point Forecast |
Lo 95 |
Hi 95 |
Point Forecast |
Lo 95 |
Hi 95 |
|
Jul |
2022 |
6169.312 |
5795.03 |
6543.59 |
1070415 |
1009574 |
1131257 |
|
Aug |
2022 |
6397.368 |
5899.45 |
6895.29 |
1120677 |
1045534 |
1195819 |
|
Sep |
2022 |
6625.425 |
6003.18 |
7247.67 |
1167325 |
1083490 |
1251160 |
|
Oct |
2022 |
6853.481 |
6103.89 |
7603.07 |
1211709 |
1120289 |
1303130 |
|
Nov |
2022 |
7081.537 |
6200.61 |
7962.46 |
1254677 |
1155010 |
1354343 |
|
Dec |
2022 |
7309.594 |
6292.92 |
8326.27 |
1296756 |
1187234 |
1406279 |
|
Jan |
2023 |
7537.65 |
6380.63 |
8694.67 |
1338280 |
1216753 |
1459807 |
|
Feb |
2023 |
7765.706 |
6463.69 |
9067.72 |
1379455 |
1243529 |
1515381 |
|
Mar |
2023 |
7993.763 |
6542.11 |
9445.41 |
1420412 |
1267658 |
1573167 |
|
Apr |
2023 |
8221.819 |
6615.94 |
9827.7 |
1461233 |
1289310 |
1633156 |
|
May |
2023 |
8449.876 |
6685.25 |
10214.5 |
1501968 |
1308693 |
1695244 |
|
Jun |
2023 |
8677.932 |
6750.11 |
10605.8 |
1542650 |
1326008 |
1759291 |
|
Jul |
2023 |
8905.988 |
6810.6 |
11001.4 |
1583298 |
1341443 |
1825152 |
|
Aug |
2023 |
9134.045 |
6866.81 |
11401.3 |
1623925 |
1355159 |
1892690 |
|
Sep |
2023 |
9362.101 |
6918.83 |
11805.4 |
1664538 |
1367293 |
1961784 |
|
Oct |
2023 |
9590.157 |
6966.72 |
12213.6 |
1705144 |
1377962 |
2032326 |
|
Nov |
2023 |
9818.214 |
7010.57 |
12625.9 |
1745744 |
1387263 |
2104226 |
|
Dec |
2023 |
10046.27 |
7050.46 |
13042.1 |
1786342 |
1395277 |
2177407 |
|
Jan |
2024 |
10274.326 |
7086.46 |
13462.2 |
1826937 |
1402074 |
2251800 |
|
Feb |
2024 |
10502.383 |
7118.63 |
13886.1 |
1867531 |
1407713 |
2327348 |
|
Mar |
2024 |
10730.439 |
7147.05 |
14313.8 |
1908124 |
1412246 |
2404001 |
|
Apr |
2024 |
10958.496 |
7171.78 |
14745.2 |
1948716 |
1415718 |
2481715 |
|
May |
2024 |
11186.552 |
7192.88 |
15180.2 |
1989309 |
1418166 |
2560451 |
|
Jun |
2024 |
11414.608 |
7210.41 |
15618.8 |
2029901 |
1419627 |
2640174 |
Figure 3: UPI payment forecasted volume
Figure 4: UPI Payment forecasted valu
Figure 4: UPI Payment forecasted value
Figure 4: UPI payment forecasted value
Conclusion:
UPI payments have risen to prominence in India's banking system because of demonetization, the increase of ICT infrastructure, rising income levels, and the spread of the pandemic. (A. & S., 2022). During the pandemic, the majority of people in the country chosen to pay their bills using UPI transactions, therefore the majority of people familiar with the system (Chakraborty et al., 2022). According to the findings of this study, there will be a generally favourable trend toward the number of UPI payments and the value of transactions during the course of the next calendar year. This is clearly indicate the upcoming challenge to the digital payment system due to huge number of transactions and related failures of the transaction. National Payment Corporation of India data shows that UPI payments were faced significant errors while sending and receiving money during transactions (UPI Report,2022). Recent studies found that most of the banks UPI payments transactions are facing technical errors during the peak time of the payments (Donda et al., 2022). Banks need to update their IT infrastructure as the growing demand from the customers to avoid the transaction failures during peak time. Some of the studies identified the key security challenges in UPI payments system where customers lost their money due to cyber frauds (Chaterji & Thomas, 2017; Chawla et al., 2021). The conclusions of this study are consistent with findings of previous studies (A & Bhat, 2021; K. Devi & Devadutta Indoria, 2021; Panse et al., 2021) that have been undertaken on the growth of digital payment systems in other countries, including India. There is a rapid growth in the use of UPI for digital transactions, large number of payments settled through the UPI platform, the trend would continue for the next calendar year. India is in the forefront of a comprehensive progress in digital payments and UPI is the harbinger in this step towards digitization of transactions. To maintain the growing trend of UPI payments in India, the government and other regulatory organizations must concentrate on the pit fall of UPI payment systems (Galhotra et al., 2021; Madwanna et al., 2021; Yogesh Chandra & Kapil, 2021) and implement remedial measures to meet the demand.
References
A, M., & Bhat, G. (2021). Digital Payment Service in India - A Case Study of Unified Payment Interface. https://doi.org/10.5281/ZENODO.5091207
A., M., & S., G. B. (2022). India’s Digital Payment Landscape – An Analysis. International Journal of Case Studies in Business, IT and Education (IJCSBE), 6(1), 223–236. https://doi.org/10.47992/IJCSBE.2581.6942.0161
Ahmed, S., & Sur, S. (2021). Change in the uses pattern of digital banking services by Indian rural MSMEs during demonetization and Covid-19 pandemic-related restrictions. Vilakshan - XIMB Journal of Management, ahead-of-print(ahead-of-print). https://doi.org/10.1108/XJM-09-2020-0138
Alsharif, M. H., Younes, M. K., & Kim, J. (2019). Time Series ARIMA Model for Prediction of Daily and Monthly Average Global Solar Radiation: The Case Study of Seoul, South Korea. Symmetry 2019, Vol. 11, Page 240, 11(2), 240. https://doi.org/10.3390/SYM11020240
Bhuyan, B., Patra, S., & Bhuian, R. K. (2021). Measurement and determinants of total factor productivity: evidence from Indian banking industry. International Journal of Productivity and Performance Management, ahead-of-print(ahead-of-print). https://doi.org/10.1108/IJPPM-06-2019-0256/FULL/XML
Chakrabarty, M., Jha, A., & Ray, P. (2020). Demonetization and digital payments in India: perception and reality. Https://Doi.Org/10.1080/13504851.2020.1752895, 28(4), 319–323. https://doi.org/10.1080/13504851.2020.1752895
Chakraborty, D., Siddiqui, A., Siddiqui, M., Rana, N. P., & Dash, G. (2022). Mobile payment apps filling value gaps: Integrating consumption values with initial trust and customer involvement. Journal of Retailing and Consumer Services, 66, 102946. https://doi.org/10.1016/J.JRETCONSER.2022.102946
Chawla, A., Manjhi, G., & Bhattacharya, G. (2021). Implications of Banking Regulations on Online Payment Failures.
Chaterji, Dr. A., & Thomas, R. (2017). Unified Payment Interface (Upi) a Catalyst Tool Supporting Digitalization – Utility, Prospects &Amp; Issues. South Dakota Review, 50(SPEC.ISSU.), 74–79. https://doi.org/10.5040/9781408169827.00000010
Chopra, R. (2017). Impact of Demonetization on Indian Economy. Global Journal of Enterprise Information System. https://doi.org/10.18311/gjeis/2017/15857
Donda, R., Arpana, M., Rajesh, D., Sagar, M. A., & Roshitha, N. (2022). Users Perceptions and Problems on Mobile Wallet Payments-A Study in Visakhapatnam City. ComFin Research, 10(1), 12–16. https://doi.org/10.34293/commerce.v10iS1-Jan.4750
Galhotra, B., Jatain, A., Bajaj, S. B., & Jaglan, V. (2021). Mobile Payments: Assessing the Threats, Challenges and Security Measures. Proceedings of the 5th International Conference on Electronics, Communication and Aerospace Technology, ICECA 2021, 997–1004. https://doi.org/10.1109/ICECA52323.2021.9676092
Gupta, R. (2021). Industry 4.0 Adaption in Indian Banking Sector—A Review and Agenda for Future Research: Https://Doi.Org/10.1177/0972262921996829. https://doi.org/10.1177/0972262921996829
Katris, C. (2021). A time series-based statistical approach for outbreak spread forecasting: Application of COVID-19 in Greece. Expert Systems with Applications, 166, 114077. https://doi.org/10.1016/J.ESWA.2020.114077
Kaur, S. J., Ali, L., Hassan, M. K., & Al-Emran, M. (2021). Adoption of digital banking channels in an emerging economy: exploring the role of in-branch efforts. Journal of Financial Services Marketing, 26(2), 107–121. https://doi.org/10.1057/S41264-020-00082-W/TABLES/3
Khandelwal, I., Adhikari, R., & Verma, G. (2015). Time Series Forecasting Using Hybrid ARIMA and ANN Models Based on DWT Decomposition. Procedia Computer Science, 48(C), 173–179. https://doi.org/10.1016/J.PROCS.2015.04.167
Kotu, V., & Deshpande, B. (2019). Time Series Forecasting. Data Science, 395–445. https://doi.org/10.1016/B978-0-12-814761-0.00012-5
Kurnia, S., Peng, F., & Liu, Y. R. (2010). Understanding the adoption of electronic banking in China. Proceedings of the Annual Hawaii International Conference on System Sciences. https://doi.org/10.1109/HICSS.2010.421
Liang, Y. H. (2009). Combining seasonal time series ARIMA method and neural networks with genetic algorithms for predicting the production value of the mechanical industry in Taiwan. Neural Computing and Applications, 18(7), 833–841. https://doi.org/10.1007/S00521-008-0216-0/TABLES/5
Liu, L., Luan, R. S., Yin, F., Zhu, X. P., & Lü, Q. (2016). Predicting the incidence of hand, foot and mouth disease in Sichuan province, China using the ARIMA model. Epidemiology & Infection, 144(1), 144–151. https://doi.org/10.1017/S0950268815001144
Madwanna, Y., Khadse, M., & Chandavarkar, B. R. (2021). Security Issues of Unified Payments Interface and Challenges: Case Study. ICSCCC 2021 - International Conference on Secure Cyber Computing and Communications, 150–154. https://doi.org/10.1109/ICSCCC51823.2021.9478078
Mishra, P., Yonar, A., Yonar, H., Kumari, B., Abotaleb, M., Das, S. S., & Patil, S. G. (2021). State of the art in total pulse production in major states of India using ARIMA techniques. Current Research in Food Science, 4, 800–806. https://doi.org/10.1016/J.CRFS.2021.10.009
Ning, Y., Kazemi, H., & Tahmasebi, P. (2022). A comparative machine learning study for time series oil production forecasting: ARIMA, LSTM, and Prophet. Computers & Geosciences, 164, 105126. https://doi.org/10.1016/J.CAGEO.2022.105126
Panse, C., Sharma, A., Bhimavarapu, V. M., Rastogi, S., & Mrudula Bhimavarapu, V. (2021). Unified Payment Interface (UPI): A Digital Innovation and Its Impact on Financial Inclusion and Economic Development. Universal Journal of Accounting and Finance, 9(3), 518–530. https://doi.org/10.13189/ujaf.2021.090326
Prasanth Shakti, S., Kamal Hassan, M., Zhenning, Y., Caytiles, R. D., & NChSN, I. (2017). Annual Automobile Sales Prediction Using ARIMA Model. International Journal of Hybrid Information Technology, 10(6), 13–22. https://doi.org/10.14257/ijhit.2017.10.6.02
Rai, K. A., Somaiya, K. J., & Rajvaidya, R. (n.d.). THE ADOPTION OF M-WALLET IN INDIA.
Rhanoui, M., Yousfi, S., Mikram, M., & Merizak, H. (2019). Forecasting financial budget time series: ARIMA random walk vs LSTM neural network. IAES International Journal of Artificial Intelligence (IJ-AI, 8(4), 317–327. https://doi.org/10.11591/ijai.v8.i4.pp317-327
Shah, H., Bhatt, V., & Shah, J. (2022). A Neoteric Technique Using ARIMA-LSTM for Time Series Analysis on Stock Market Forecasting. 381–392. https://doi.org/10.1007/978-981-16-5952-2_33
Shree, S., Pratap, B., Saroy, R., & Dhal, S. (2021). Digital payments and consumer experience in India: a survey based empirical study. Journal of Banking and Financial Technology 2021 5:1, 5(1), 1–20. https://doi.org/10.1007/S42786-020-00024-Z
Sinha, M., Majra, H., Hutchins, J., & Saxena, R. (2019). Mobile payments in India: the privacy factor. International Journal of Bank Marketing. https://doi.org/10.1108/IJBM-05-2017-0099
Usmonova Durdona Shukhratovna. (2021). Digital banking in the eyes of young professionals | Academic Journal of Digital Economics and Stability. Academic Journal of Digital Economics and Stability, 6, 163–167. http://economics.academicjournal.io/index.php/economics/article/view/91
Yogesh Chandra, V., & Kapil, B. (2021). Critical Study of Unified Payment Interface (UPI): E-Payment Mode of Digital Revolution | Academic Social Research:(P) ISSN: 2456-2645 , Impact Factor: 4.928 (UGC APPROVED 47715). Academic Social Research, 7(4). http://asr.academicsocialresearch.co.in/index.php/ASR/article/view/582
