
Forecasting earnings for publicly traded companies is of paramount significance for investments, which is the background of this research. This holds particularly true in emerging markets where the coverage of these companies by financial analysts' predictions is limited. This research investigation delves into the prediction inaccuracies of cutting-edge time series forecasting algorithms created by major technology companies such as Facebook, LinkedIn, Amazon, and Google. These techniques are employed to analyze earnings per share data for publicly traded Polish companies during the period spanning from the financial crisis to the pandemic shock. My objective was to compare prediction errors of analyzed models, using scientifically defined error measures and a series of statistical tests. The seasonal random walk model demonstrated the lowest error of prediction, which might be attributable to the overfitting of complex models.
Citation: Wojciech Kuryłek. Can we profit from BigTechs' time series models in predicting earnings per share? Evidence from Poland[J]. Data Science in Finance and Economics, 2024, 4(2): 218-235. doi: 10.3934/DSFE.2024008
[1] | Syeda Nadiah Fatima Nahri, Shengzhi Du, Barend J. van Wyk, Oluwaseun Kayode Ajayi . A comparative study on time-delay estimation for time-delay nonlinear system control. AIMS Electronics and Electrical Engineering, 2025, 9(3): 314-338. doi: 10.3934/electreng.2025015 |
[2] | José M. Campos-Salazar, Pablo Lecaros, Rodrigo Sandoval-García . Dynamic analysis and comparison of the performance of linear and nonlinear controllers applied to a nonlinear non-interactive and interactive process. AIMS Electronics and Electrical Engineering, 2024, 8(4): 441-465. doi: 10.3934/electreng.2024021 |
[3] | José M. Campos-Salazar, Roya Rafiezadeh, Felipe Santander, Juan L. Aguayo-Lazcano, Nicolás Kunakov . Comprehensive GSSA and D-Q frame dynamic modeling of dual-active-bridge DC-DC converters. AIMS Electronics and Electrical Engineering, 2025, 9(3): 288-313. doi: 10.3934/electreng.2025014 |
[4] | Jun Yoneyama . H∞ disturbance attenuation of nonlinear networked control systems via Takagi-Sugeno fuzzy model. AIMS Electronics and Electrical Engineering, 2019, 3(3): 257-273. doi: 10.3934/ElectrEng.2019.3.257 |
[5] | Thanh Tung Pham, Chi-Ngon Nguyen . Adaptive PID sliding mode control based on new Quasi-sliding mode and radial basis function neural network for Omni-directional mobile robot. AIMS Electronics and Electrical Engineering, 2023, 7(2): 121-134. doi: 10.3934/electreng.2023007 |
[6] | José M. Campos-Salazar, Juan L. Aguayo-Lazcano, Roya Rafiezadeh . Non-ideal two-level battery charger—modeling and simulation. AIMS Electronics and Electrical Engineering, 2025, 9(1): 60-80. doi: 10.3934/electreng.2025004 |
[7] | Qichun Zhang, Xuewu Dai . Special Issue "Uncertainties in large-scale networked control systems". AIMS Electronics and Electrical Engineering, 2020, 4(4): 345-346. doi: 10.3934/ElectrEng.2020.4.345 |
[8] | Yanming Wu, Zelun Wang, Guanglei Meng, Jinguo Liu . Neural networks-based event-triggered consensus control for nonlinear multiagent systems with communication link faults and DoS attacks. AIMS Electronics and Electrical Engineering, 2024, 8(3): 332-349. doi: 10.3934/electreng.2024015 |
[9] | Rasool M. Imran, Kadhim Hamzah Chalok, Siraj A. M. Nasrallah . Innovative two-stage thermal control of DC-DC converter for hybrid PV-battery system. AIMS Electronics and Electrical Engineering, 2025, 9(1): 26-45. doi: 10.3934/electreng.2025002 |
[10] | Qichun Zhang . Performance enhanced Kalman filter design for non-Gaussian stochastic systems with data-based minimum entropy optimisation. AIMS Electronics and Electrical Engineering, 2019, 3(4): 382-396. doi: 10.3934/ElectrEng.2019.4.382 |
Forecasting earnings for publicly traded companies is of paramount significance for investments, which is the background of this research. This holds particularly true in emerging markets where the coverage of these companies by financial analysts' predictions is limited. This research investigation delves into the prediction inaccuracies of cutting-edge time series forecasting algorithms created by major technology companies such as Facebook, LinkedIn, Amazon, and Google. These techniques are employed to analyze earnings per share data for publicly traded Polish companies during the period spanning from the financial crisis to the pandemic shock. My objective was to compare prediction errors of analyzed models, using scientifically defined error measures and a series of statistical tests. The seasonal random walk model demonstrated the lowest error of prediction, which might be attributable to the overfitting of complex models.
This work is motivated by the networked control system (NCS) or remote control system where the media that connects the controller and actuator causes random packet delays or dropouts. This type of control systems are anticipated to have vast applications. The readers are referred to the review paper [1]. The common features of NCS are random packet delays or dropouts, competition of multiple nodes in the network, data quantization, etc., and this makes the modeling and stability analysis very challenging. Markovian type regime switching models have been proved to be very successful in modeling NCS [2,3]. However, the stability of such systems still remains a challenge. Most stability criteria are given as complicated LMI conditions [4,5] via Lyapunov function approach, or dwell time [6], or delay Riccati equations [7]. A stability condition using hybrid system analysis is found in [8], which is again a sufficient condition and is not easy to verify. A neat sufficient and necessary condition is in great need.
In order to overcome the constraints of NCS, such as random packet delays or dropouts, it is a natural idea to estimate system state and compensate for packet disorders [9]. One believes that in this way better system performance and better system stability can be achieved. Predictive control has been effectively applied to NCS [10,11,12,13]. This control method generates a sequence of future control variables, which can be used to compensate for packet disorder or dropouts [14], and to estimate unknown system state as well [15,16]. However, as mentioned earlier, the stability conditions are usually hard to verify. To analyze the stability issue, one must have a thorough understanding of the dynamics of NCS with random time delays and state estimate.
In this paper we propose a predictive control method of NCS with random packet delays and state estimate/compensation, and clearly show the structure of this system. This control method has much less computational burden compared to classical predictive control method. Then the system is modeled as a regime switching system, and an upper bound on the number of regimes is provided. By the word ``scale'' we mean the number of distinct regimes (denoted by
Throughout this paper we use
Unlike most NCSs where the objective systems on the actuator side are time invariant, here we assume that the objective system is a regime switching system with Markovian jumps, and we call it a Markov Jump System (MJS). The application of switching systems in engineering can be found, for example, in [17,18,19]. We let
We begin with simple and practically reasonable assumptions on this system.
● The actuator receives signals with random delays.
The actuator receives control signals from the controller and performs the required actions. Due to the random packet delays, the actuator might receive no signal, or receive a packet that is randomly delayed, or receive multiple packets. If more than one packet arrive at the same time, the actuator executes the most recent one.
● Obsolete signals are discarded.
If, at a certain time instant, a control variable has been applied, then any control signal older than it but arrives at later times will be discarded. For example, if
● The delay times are bounded.
The backward time delay
● The controller sends out signals at every discrete instant.
The controller receives state feedback from the actuator via randomly delayed packets, and the controller estimates the future system state and sends out a new control signal. If no state feedback is received at a certain time instant, the controller simply sends a zero control signal.
● Past control signals are recorded.
The controller keeps a record of past control signals that have been sent. This information is used to calculate a new control signal whenever a feedback packet arrives.
● Packet size is limited.
The network bandwidth is very limited so that each time the controller and actuator just send out one small packet to the other party. That means, sending out a large packet containing a long sequence of controls or system states is not applicable. (Otherwise this system reduces to classical control system without time delay.)
The controlled system on the actuator side becomes
xk+1=a(ik)xk+b(j)(ik)uk−τ2, | (1.1) |
where
The NCS is a special dynamical system that might be easy to describe in words, but is hard to understand because its dynamics looks chaotic. One must get down to the bottom and have a careful dissemination to obtain a clear understanding. In what follows we shall do this work.
We use the triplet
If no control signal arrives at time
The expression
We use a similar mechanism to represent the situation at the controller side. The pair
Notice that the received information depends on the sequence of the past states
Proposition 2.1. An upper bound on the number of different regimes of this problem (denoted by
m2+T1⋅(1+T2)2+T1⋅(2+T2)/2⋅(1+T1)⋅(2+T1)/2 | (2.1) |
Proof. On the controller side we consider the pair
Similarly, on the actuator side, any pair
Due to the feedback delay, a sequence with length
The controller then calculates and sends out a new control variable
However, recall that not all the transitions
m⋅(1+T2)⋅(2+T2)/2⋅[m⋅(1+T2)]1+T1=m2+T1⋅(1+T2)2+T1⋅(2+T2)/2. |
Now the proof is complete.
It is clear that to represent the complete system scenario, we need
In each scenario in
zk=[xk(ik),xk−1(ik−1),...,xk−T1(ik−T1),uk−1,uk−2,...,uk−T1−T2]T, | (3.1) |
where the superscript
zk+1=A(C(k)2,τ(k)2,ik)zk+B(C(k)2,τ(k)2,ik)uk, | (3.2) |
which is an extension of (1.1). Since each entry in (3.1) is necessary to describe the system dynamics in a certain regime, we are clear that
If
A(C(k)2,τ(k)2,ik)=(a(ik) 0 0 0 0⋯1 0 0 0 0⋯0 1 0 0 0⋯⋯⋯⋯⋯⋯⋯0(∗)⋯ 0 0 0⋯0⋯ 0 1 0⋯⋯⋯⋯⋯⋯⋯),B(C(k)2,τ(k)2,ik)=[0 0 0⋯ 1(∗) 0 0⋯]T, |
where
If
If
A(0,τ(k)2,ik)=(a(ik) 0⋯b(ik) 0 0⋯1 0 0 0 0 0⋯0 1 0 0 0 0⋯⋯⋯⋯⋯⋯⋯⋯0(∗)⋯ 0 0 0 0⋯0⋯ 0 1 0 0⋯⋯⋯⋯⋯⋯⋯⋯),B(0,τ(k)2,ik)=[0 0 0⋯ 1(∗) 0 0⋯]T, |
where the position of
On the controller side we have
uk=−F(C(k)1,τ(k)1,C(k−τ(k)1−1)2,τ(k−τ(k)1−1)2,ik−τ(k)1−1)zk. | (3.3) |
If
uk−τ(k)1−τ(k−τ(k)1−1)2,uk−τ(k)1−τ(k−τ(k)1−1)2+1,...,uk−1, | (3.4) |
the controller is able to estimate the system state
Recall that if the controller does not receive any feedback packet at present time
We need to point out that there is much flexibility in calculating the new control variable, which might be the most resourceful research in NCS. In this paper we present our control method as follows. Firstly, based on the most recent information
ˆxk−τ(k)1+1=a(ˆik−τ(k)1)xk−τ(k)1+b(ˆik−τ(k)1)uk−τ(k)1−τ(k−τ(k)1−1)2, | (3.5) |
and
ˆxk−τ(k)1+2=a(ˆik−τ(k)1+1)ˆxk−τ(k)1+1+b(ˆik−τ(k)1+1)uk−τ(k)1−τ(k−τ(k)1−1)2+1, | (3.6) |
and so forth until
One may notice that the system mode information
Based on the probability transition matrix
To calculate a new control
ˆik−τ(k)1,ˆik−τ(k)1+1,⋯,ˆik+τ(k−τ(k)1−1)2−1,ˆik+τ(k−τ(k)1−1)2,..,ˆik−τ(k)1+L−1, | (3.7) |
where
J(ˆxk+τ2,ˆik+τ(k−τ(k)1−1)2−1,τ(k)1,τ(k−τ(k)1−1)2)=L−τ(k)1∑j=τ(k−τ(k)1−1)2+1ˆxTk+jQ(k+j)ˆxk+j+L−τ(k)1−τ(k−τ(k)1−1)2−1∑j=0ˆuTk+jR(k+j)ˆuk+j, | (3.8) |
where
Unlike the usual cost function which is the expected value along all possible future paths, this cost function (3.8) is calculated along the deterministic path (3.7), and because the optimization is performed on one path, a huge amount of calculation is saved. To see this in more details, let us consider a control method which chooses future controls
J(ˆxk+τ2,ˆik+τ(k−τ(k)1−1)2−1,τ(k)1,τ(k−τ(k)1−1)2)=L−τ(k)1∑j=τ(k−τ(k)1−1)2+1E[ˆxTk+jQ(k+j)ˆxk+j|xk−τ(k)1]+L−τ(k)1−τ(k−τ(k)1−1)2−1∑j=0ˆuTk+jR(k+j)ˆuk+j, | (3.9) |
where
But for our proposed simplified predictive control method, a natural question remains: what if the path we predicted does not actually occur? Remember that so far we have not used rolling optimization which is the true power of RHC. The calculation of future control sequences is repeated at each sample instant, and if a prediction error occurs, it will be corrected by re-planning at the next sample instant.
Actually we can make this calculation faster by noticing that the control format in RHC with quadratic criterion function (3.8) is given by linear feedback form, see, e.g., [20]. We borrow the iterative algorithm in [20] for the calculation of state feedback control in time varying case, then we obtain
uk=−F(ˆik+τ(k−τ(k)1−1)2−1,τ(k)1,τ(k−τ(k)1−1)2)ˆxk+τ(k−τ(k)1−1)2. | (3.10) |
Recall that
Dk−τ(k)1+1=(0 ⋯0 a(ˆik−τ(k)1) 0 ⋯0 b(ˆik−τ(k)1)⋯0 ⋯⋯ ⋯ ⋯ 0 1 0⋯0 ⋯⋯ ⋯⋯ 1 0 0⋯⋯⋯⋯ ⋯⋯⋯⋯ ⋯⋯), |
where
Dk−τ(k)1+1zk=(ˆxk−τ(k)1+1 uk−τ(k)1−τ(k−τ(k)1−1)2+1⋯⋯uk−1)T |
by (3.5).
Now we define
Dk−τ(k)1+2=(a(ˆik−τ(k)1+1) b(ˆik−τ(k)1+1)0 0⋯0 01 0⋯0 00 1⋯⋯ ⋯⋯ ⋯⋯), |
and get
Dk−τ(k)1+2Dk−τ(k)1+1zk=(ˆxk−τ(k)1+2 uk−τ(k)1−τ(k−τ(k)1−1)2+2 ⋯uk−1)T |
by (3.6). Continue this process until we reach
Dk+τ(k−τ(k)1−1)2⋯Dk−τ(k)1+1zk=ˆxk+τ(k−τ(k)1−1)2. | (3.11) |
Substituting (3.11) in (3.10) yields
uk=−F(ˆik+τ(k−τ(k)1−1)2−1,τ(k)1,τ(k−τ(k)1−1)2)⋅Dk+τ(k−τ(k)1−1)2⋯Dk−τ(k)1+1zk. |
Since
F(C(k)1,τ(k)1,C(k−τ(k)1−1)2,τ(k−τ(k)1−1)2,ik−τ(k)1−1)=F(ˆik+τ(k−τ(k)1−1)2−1,τ(k)1,τ(k−τ(k)1−1)2)⋅Dk+τ(k−τ(k)1−1)2⋯Dk−τ(k)1+1. | (3.12) |
In the next section we shall investigate the stability of this control algorithm.
Putting (3.3) in (3.2) yields
zk+1=A(C(k)2,τ(k)2,ik)zk−B(C(k)2,τ(k)2,ik)⋅F(C(k)1,τ(k)1,C(k−τ(k)1−1)2,τ(k−τ(k)1−1)2,ik−τ(k)1−1)zk. | (4.1) |
Denote
H=H(C(k)2,τ(k)2,ik,C(k)1,τ(k)1,C(k−τ(k)1−1)2,τ(k−τ(k)1−1)2,ik−τ(k)1−1)=A(C(k)2,τ(k)2,ik)−B(C(k)2,τ(k)2,ik)⋅F(C(k)1,τ(k)1,C(k−τ(k)1−1)2,τ(k−τ(k)1−1)2,ik−τ(k)1−1), |
then we obtain the following dynamics
zk+1=H⋅zk, | (4.2) |
where
The usual stability criteria on MJSs are mean convergence (MC) and mean square convergence (MSC). We say that a system is uniformly MC if
F=diag(H)⋅(P′⊗Id),A=diag(H⊗H)⋅(P′⊗Id2), | (4.3) |
where
Theorem 4.1. The MJS (4.1) is uniformly MC if and only if
Proof. The proof can be found in [22]. Here we just point out that the matrices
In the numerical example we shall test both stabilities.
Remark 4.1. We are able to provide this sufficient and necessary condition on stability, and this is largely due to the fact that the structure of this NCS has been clearly revealed.
For the complete representation
(C(k)2,τ(k)2,ik,C(k)1,τ(k)1)→(C(k+1)2,τ(k+1)2,ik+1,C(k+1)1,τ(k+1)1). | (5.1) |
As we mentioned earlier, the regimes do not have full reachability, but the transition (5.1) can be decomposed into three parts that are treated independently. The first part
We consider an example where the objective system, which is a MJS, has two modes (
[x(1)k+1x(2)k+1]=[1.20.2 10 ][x(1)kx(2)k]+[0.05 0.1]uk−τ2, |
and
[x(1)k+1x(2)k+1]=[0.70.2 10 ][x(1)kx(2)k]+[0.10.3]uk−τ2, |
respectively. This system is controlled by a controller through a network with random time delays. It is assumed that
The matrix
Pm=(0.8 0.20.2 0.8). |
To construct
(0,0)(0,1) (0,2)(1,0) (1,1)(2,0)(0,0) 0.70 00.3 00 (0,1) 0.40.4 00 0.20 (0,2) 0.50.2 0.30 00 (1,0) 0.30.4 00 00.3 (1,1) 0.50.2 0.30 00 (2,0) 0.40.2 0.40 00 | (6.1) |
Now that
P1=(0.6 0 0.40.5 0.5 00.7 0.3 0). |
By the derivation in this paper, we need to keep
(0,0)→(0,0)→(0,0)(0,0)→(0,0)→(1,0)(0,0)→(1,0)→(0,0)(0,0)→(1,0)→(0,1)(0,0)→(1,0)→(2,0)⋯⋯⋯, |
and the total number of all possible 3-pair paths is 45. The estimate from (2.1) on this part is given by
The vector
Without any control we have
We choose the prediction horizon
The MJS stability certainly depends on the transition probabilities among the regimes. For instance, if we use
Pm=(0.85 0.150.25 0.75),P1=(0.7 0 0.30.6 0.4 00.8 0.2 0), |
and
P2=(0.8 0 0 0.2 0 00.5 0.3 0 00.2 00.6 0.3 0.1 00 00.4 0.30 00 0.30.5 0.20.3 00 00.7 0.20.1 00 0), |
then again, without control we have
Then we apply the proposed predictive control method and have
To overcome the constraints of NCS such as random packet delays/disorders, it is a natural idea to estimate the system states and send out a control signal that compensates the time delay, hence the expectation of a better control performance. In this paper we modeled the NCS as a regime switching system and proposed a simplified predictive control method. The regime estimate (2.1), which is one of the main contributions of this paper, illustrates that this seemingly small system actually has a very large scale. The structure of this system has been clearly revealed, and a concise sufficient and necessary condition on stability is obtained. Numerical examples clearly show the features of this dynamical system.
This work is supported by Barrios Technology Faculty Fellowship from University of Houston - Clear Lake, 2017-2018.
The author declares no conflict of interest in this paper.
[1] |
Ahmadpour A, Etemadi H, Moshashaei S, et al. (2015) Earnings Per Share Forecast Using Extracted Rules from Trained Neural Network by Genetic Algorithm. Comput Econ 46: 55–63. https://doi.org/10.1007/s10614-014-9455-6 doi: 10.1007/s10614-014-9455-6
![]() |
[2] |
Aidan IA, Al-Jeznawi D, Al-Zwainy FM, et al. (2020) Predicting earned value indexes in residential complexes' construction projects using artificial neural network model. Int J Intell Eng Syst 13: 248–259. http://dx.doi.org/10.22266/ijies2020.0831.22 doi: 10.22266/ijies2020.0831.22
![]() |
[3] |
Alexander RA, Govern DM (1994) A New and Simpler Approximation for ANOVA under Variance Heterogeneity. J Educ Stat 19: 91–101. http://dx.doi.org/10.2307/1165140 doi: 10.2307/1165140
![]() |
[4] | Ahammad P, Al Orjany SE, Chen A, et al. (2022) Greykite: deploying flexible forecasting at scale at LinkedIn. Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 3007–3017. https://doi.org/10.1145/3534678.3539165 |
[5] |
Al-Somaydaii JA, Al-Zwainy FM, Hammoody O, et al. (2022) Forecasting and determining of cost performance index of tunnels projects using artificial neural networks. Int J Comput Civil Struct Eng 18: 51–60. https://doi.org/10.22337/2587-9618-2022-18-1-51-60 doi: 10.22337/2587-9618-2022-18-1-51-60
![]() |
[6] |
Al-Zwainy FM, Amer R, Khaleel T, et al. (2016) Reviewing of the simulation models in cost management of the construction projects. Civil Eng J 2: 607–622. https://doi.org/10.28991/cej-2016-00000063 doi: 10.28991/cej-2016-00000063
![]() |
[7] | AL-Zwainy FM, Huda F, Ibrahim A, et al. (2018) Development of an Analytical Software for Cost Estimation for Highway Project. Int J Appl Eng Res 13: 6944–6951. |
[8] |
Al-Zwainy FM, Jaber FK, Jasim NA, et al. (2020) Forecasting techniques in construction industry: earned value indicators and performance models. Sci Rev Eng Environ Sci 29: 234–243. https://doi.org/10.22630/PNIKS.2020.29.2.20 doi: 10.22630/PNIKS.2020.29.2.20
![]() |
[9] | Al-Zwainy FM, Raheem SH (2020) Innovation of Analytical Software for Financing Construction Projects: Infrastructure Projects in Iraq as a Case Study, In: IOP Conference Series: Materials Science and Engineering, IOP Publishing, 978: 012015. https://doi.org/10.1088/1757-899X/978/1/012015 |
[10] |
Arık SÖ, Lim B, Loeff N, et al. (2021) Temporal Fusion Transformers for interpretable multi-horizon time series forecasting. Int J Forecast 37: 1748–1764. https://doi.org/10.1016/j.ijforecast.2021.03.012 doi: 10.1016/j.ijforecast.2021.03.012
![]() |
[11] |
Ball R, Watts R (1972) Some Time Series Properties of Accounting Income. J Financ 27: 663–681. http://dx.doi.org/10.1111/j.1540-6261.1972.tb00991.x doi: 10.1111/j.1540-6261.1972.tb00991.x
![]() |
[12] | Bathke Jr AW, Lorek KS (1984) The Relationship between Time-Series Models and the Security Market's Expectation of Quarterly Earnings. Account Rev 59: 163–176. |
[13] | Bengio Y, Courville A, Goodfellow I, et al. (2017) Deep Learning. Cambridge, Massachusetts: The MIT Press. |
[14] |
Bradshaw M, Drake M, Myers J, et al. (2012) A re-examination of analysts' superiority over time-series forecasts of annual earnings. Rev Account Stud 17: 944–968. http://dx.doi.org/10.1007/s11142-012-9185-8 doi: 10.1007/s11142-012-9185-8
![]() |
[15] | Brandon Ch, Jarrett JE, Khumawala SB, et al. (1983) On the predictability of corporate earnings per share. J Bus Financ Account 10: 373–387. |
[16] |
Brandon Ch, Jarrett JE, Khumawala SB, et al. (1986) Comparing forecast accuracy for exponential smoothing models of earnings-per-share data for financial decision making. Decision Sci 17: 186–194. http://dx.doi.org/10.1111/j.1540-5915.1986.tb00220.x doi: 10.1111/j.1540-5915.1986.tb00220.x
![]() |
[17] |
Brandon Ch, Jarrett JE, Khumawala SB, et al. (1987) A Comparative Study of the Forecasting Accuracy of Holt‐Winters and Economic Indicator Models of Earnings Per Share For Financial Decision Making. Manag Financ 13: 10–15. http://dx.doi.org/10.1108/eb013581 doi: 10.1108/eb013581
![]() |
[18] |
Brooks LD, Buckmaster DA (1976) Further Evidence of The Time Series Properties Of Accounting Income. J Financ 31: 1359–1373. http://dx.doi.org/10.1111/j.1540-6261.1976.tb03218.x doi: 10.1111/j.1540-6261.1976.tb03218.x
![]() |
[19] |
Brown LD, Griffin PA, Hagerman RL, et al. (1987) Security analyst superiority relative to univariate time-series models in forecasting quarterly earnings. J Account Econ 9: 61–87. http://dx.doi.org/10.1016/0165-4101(87)90017-6 doi: 10.1016/0165-4101(87)90017-6
![]() |
[20] |
Brown LD, Rozeff MS (1979) Univariate Time-Series Models of Quarterly Accounting Earnings per Share: A Proposed Model. J Account Res 17: 179–189. http://dx.doi.org/10.2307/2490312 doi: 10.2307/2490312
![]() |
[21] | Cao Q, Gan Q (2009) Forecasting EPS of Chinese listed companies using a neural network with genetic algorithm. 15th Americas Conference on Information Systems 2009, AMCIS 2009, 2791–2981. |
[22] |
Cao Q, Parry M (2009) Neural network earnings per share forecasting models: A comparison of backward propagation and the genetic algorithm. Decis Support Syst 47: 32–41. http://dx.doi.org/10.1016/j.dss.2008.12.011 doi: 10.1016/j.dss.2008.12.011
![]() |
[23] |
Cao Q, Schniederjans MJ, Zhang W, et al. (2004) Neural network earnings per share forecasting models: A comparative analysis of alternative methods. Decision Sci 35: 205–237. https://doi.org/10.1111/j.00117315.2004.02674.x doi: 10.1111/j.00117315.2004.02674.x
![]() |
[24] |
Conroy R, Harris T (1987) Consensus Forecasts of Corporate Earnings: Analysts' Forecasts and Time Series Methods. Manag Sci 33: 725–738. http://dx.doi.org/10.1287/mnsc.33.6.725 doi: 10.1287/mnsc.33.6.725
![]() |
[25] | Corder GW, Foreman DI (2009) Comparing More than Two Unrelated Samples: The Kruskal–Wallis H-Test. In: Nonparametric Statistics for Non-Statisticians, John Wiley & Sons, Hoboken, New Jersey, 99–121. http://dx.doi.org/10.1002/9781118165881 |
[26] |
Dreher S, Eichfelder S, Noth F, et al. (2024) Does IFRS information on tax loss carryforwards and negative performance improve predictions of earnings and cash flows? J Bus Econ 94: 1–39. http://dx.doi.org/10.1007/s11573-023-01147-7 doi: 10.1007/s11573-023-01147-7
![]() |
[27] | Elend L, Kramer O, Lopatta J, et al. (2020) Earnings prediction with deep learning. German Conference on Artificial Intelligence (Künstliche Intelligenz), KI 2020: Advances in Artificial Intelligence, 267–274. http://dx.doi.org/10.1007/978-3-030-58285-2_22 |
[28] |
Elton EJ, Gruber MJ (1972) Earnings Estimates and the Accuracy of Expectational Data. Manag Sci 18: B409–B424. http://dx.doi.org/10.1287/mnsc.18.8.B409 doi: 10.1287/mnsc.18.8.B409
![]() |
[29] |
Finger CA (1994) The ability of earnings to predict future earnings and cash flow. J Account Res 32: 210–223. http://dx.doi.org/10.2307/2491282 doi: 10.2307/2491282
![]() |
[30] |
Flagmeier V (2022) The information content of deferred taxes under IFRS. Eu Account Rev 31: 495–518. http://dx.doi.org/10.1080/09638180.2020.1826338 doi: 10.1080/09638180.2020.1826338
![]() |
[31] | Foster G (1977) Quarterly Accounting Data: Time-Series Properties and Predictive-Ability Results. Account Rev 52: 1–21. |
[32] |
Flunkert V, Gasthaus J, Januschowski T, et al. (2020) DeepAR: Probabilistic forecasting with autoregressive recurrent networks. Int J Forecast 36: 1181–1191. https://doi.org/10.1016/j.ijforecast.2019.07.001 doi: 10.1016/j.ijforecast.2019.07.001
![]() |
[33] |
Gaio L, Gatsios R, Lima F, et al. (2021) Re-examining analyst superiority in forecasting results of publicly-traded Brazilian companies. Revista de Administracao Mackenzie 22: eRAMF210164. https://doi.org/10.1590/1678-6971/eramf210164 doi: 10.1590/1678-6971/eramf210164
![]() |
[34] |
Gerakos J, Gramacy RB (2013) Regression-Based Earnings Forecasts. Chicago Booth Research Paper, 12–26. https://doi.org/10.2139/ssrn.2112137 doi: 10.2139/ssrn.2112137
![]() |
[35] |
Griffin P (1977) The Time-Series Behavior of Quarterly Earnings: Preliminary Evidence. J Account Res 15: 71–83. http://dx.doi.org/10.2307/2490556 doi: 10.2307/2490556
![]() |
[36] | Grigaliūnienė Ž (2013) Time-series models forecasting performance in the Baltic stock market. Organ Market Emerg E 4: 104–120. |
[37] | Holt Ch C (1957) Forecasting seasonals and trends by exponentially weighted moving averages. Working Paper, Carnegie Institute of Technology. |
[38] |
Holt Ch C (2004) Forecasting seasonals and trends by exponentially weighted moving averages. J Econ Soc Meas 29: 123–125. https://doi.org/10.1016/j.ijforecast.2003.09.015 doi: 10.1016/j.ijforecast.2003.09.015
![]() |
[39] |
Jarrett JE (2008) Evaluating Methods for Forecasting Earnings Per Share. Manag Financ 16: 30–35. http://dx.doi.org/10.1108/eb013647 doi: 10.1108/eb013647
![]() |
[40] |
Johnson TE, Schmitt TG (1974) Effectiveness of Earnings Per Share Forecasts. Financ Manag 3: 64–72. http://dx.doi.org/10.2307/3665292 doi: 10.2307/3665292
![]() |
[41] |
Kim S, Kim H (2016) A new metric of absolute percentage error for intermittent demand forecasts. Int J Forecast 32: 669–679. http://dx.doi.org/10.1016/j.ijforecast.2015.12.003 doi: 10.1016/j.ijforecast.2015.12.003
![]() |
[42] |
Kuryłek W (2023a) The modeling of earnings per share of Polish companies for the post-financial crisis period using random walk and ARIMA models. J Bank Financ Econ 1: 26–43. http://dx.doi.org/10.7172/2353-6845.jbfe.2023.1.2 doi: 10.7172/2353-6845.jbfe.2023.1.2
![]() |
[43] | Kuryłek W (2023b) Can exponential smoothing do better than seasonal random walk for earnings per share forecasting in Poland? Bank Credit 54: 651–672. |
[44] | Lacina M, Lee B, Xu R, et al. (2011) An evaluation of financial analysts and naï ve methods in forecasting long-term earnings, In: K. D Lawrence & R. K. Klimberg (Eds.), Advances in business and management forecasting, Bingley, UK: Emerald, 77–101. http://dx.doi.org/10.1108/S1477-4070(2011)0000008009 |
[45] |
Lai S, Li H (2006) The predictive power of quarterly earnings per share based on time series and artificial intelligence model. Appl Financ Econ 16: 1375–1388. http://dx.doi.org/10.1080/09603100600592752 doi: 10.1080/09603100600592752
![]() |
[46] |
Letham B, Taylor SJ (2018) Forecasting at scale. Am Stat 72: 37–45. https://doi.org/10.7287/peerj.preprints.3190v1 doi: 10.7287/peerj.preprints.3190v1
![]() |
[47] |
Lev B, Souginannis T (2010) The usefulness of accounting estimates for predicting cash flows and earnings. Rev Account Stud 15: 779–807. http://dx.doi.org/10.1007/s11142-009-9107-6 doi: 10.1007/s11142-009-9107-6
![]() |
[48] |
Lorek KS (1979) Predicting Annual Net Earnings with Quarterly Earnings Time-Series Models. J Account Res 17: 190–204. http://dx.doi.org/10.2307/2490313 doi: 10.2307/2490313
![]() |
[49] | Lorek KS, Willinger GL (1996) A multivariate time-series model for cash-flow data. Account Rev 71: 81–101. |
[50] |
Lorek KS, Willinger GL (2007) The contextual nature of the predictive power of statistically-based quarterly earnings models. Rev Q Financ Account 28: 1–22. http://dx.doi.org/10.1007/s11156-006-0001-z doi: 10.1007/s11156-006-0001-z
![]() |
[51] | Lowry R (2014) Concepts and Applications of Inferential Statistics, Chapter 14. Available from: https://onlinebooks.lib, rary.upenn.edu/webbin/book/lookupid?key = olbp66608. |
[52] |
Pagach DP, Warr RS (2020) Analysts versus time-series forecasts of quarterly earnings: A maintained hypothesis revisited. Adv Account 51: 1–15. http://dx.doi.org/10.1016/j.adiac.2020.100497 doi: 10.1016/j.adiac.2020.100497
![]() |
[53] |
Ruland W (1980) On the Choice of Simple Extrapolative Model Forecasts of Annual Earnings. Financ Manag 9: 30–37. http://dx.doi.org/10.2307/3665165 doi: 10.2307/3665165
![]() |
[54] | Watts RL (1975) The Time Series Behavior of Quarterly Earnings. Working Paper, Department of Commerce, University of New Castle. |
[55] |
Wilcoxon F (1945) Individual comparisons by ranking methods. Biometrics Bulletin 1: 80–83. http://dx.doi.org/10.2307/3001968 doi: 10.2307/3001968
![]() |
[56] | Xiaoqiang W (2022) Research on enterprise financial performance evaluation method based on data mining. In: 2022 IEEE 2nd International Conference on Electronic Technology, Communication and Information (ICETCI). https://doi.org/10.1109/icetci55101.2022.9832404 |