Hindawi

Mathematical Problems in Engineering

Volume 2018, Article ID 9280590, 12 pages
https://doi.org/10.1155/2018/9280590

Research Article
Predicting Stock Price Trend Using MACD Optimized by
Historical Volatility

Jian Wang and Junseok Kim

Department of Mathematics, Korea University, Seoul 02841, Republic of Korea

Correspondence should be addressed to Junseok Kim; [email protected]

Received 18 September 2018; Revised 13 November 2018; Accepted 21 November 2018; Published 25 December 2018

Academic Editor: Luis Martı́nez

Copyright © 2018 Jian Wang and Junseok Kim. This is an open access article distributed under the Creative Commons Attribution
License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly
cited.

With the rapid development of the financial market, many professional traders use technical indicators to analyze the stock
market. As one of these technical indicators, moving average convergence divergence (MACD) is widely applied by many investors.
MACD is a momentum indicator derived from the exponential moving average (EMA) or exponentially weighted moving average
(EWMA), which reacts more significantly to recent price changes than the simple moving average (SMA). Traders find the analysis
of 12- and 26-day EMA very useful and insightful for determining buy-and-sell points. The purpose of this study is to develop an
effective method for predicting the stock price trend. Typically, the traditional EMA is calculated using a fixed weight; however,
in this study, we use a changing weight based on the historical volatility. We denote the historical volatility index as HVIX and
the new MACD as MACD-HVIX. We test the stability of MACD-HVIX and compare it with that of MACD. Furthermore, the
validity of the MACD-HVIX index is tested by using the trend recognition accuracy. We compare the accuracy between a MACD
histogram and a MACD-HVIX histogram and find that the accuracy of using MACD-HVIX histogram is 55.55% higher than that
of the MACD histogram when we use the buy-and-sell strategy. When we use the buy-and-hold strategy for 5 and 10 days, the
prediction accuracy of MACD-HVIX is 33.33% and 12% higher than that of the traditional MACD strategy, respectively. We found
that the new indicator is more stable. Therefore, the improved stock price forecasting model can predict the trend of stock prices
and help investors augment their return in the stock market.

1. Introduction predicted the minute-ahead stock price by using singular

spectrum analysis and support vector regression. Researchers
Securities investment is a financial activity influenced by have also used other methods to forecast stock markets.
many factors such as politics, economy, and psychology of Singh et al. [7] designed a forecasting model consisting of
investors. Its process of change is nonlinear and multifractal fuzzy theory and particle swarm optimization to predict
[1]. The stock market has high-risk characteristics; i.e., if the stock markets using historical data from the State Bank of
stock price volatility is excessive or the stability is low, the risk India. Lahmiri et al. [8] proposed an intelligent ensemble
is uncontrollable. Financial asset returns in the short term are forecasting system for stock market fluctuations based on
persistent; however, those in the long term will be reversed symmetric and asymmetric wavelet functions. Das et al. [9]
[2]. proposed a hybridized machine-learning framework using
Asness [3] reported that the stock, foreign exchange, a self-adaptive multipopulation-based Jaya algorithm for
and commodity markets have a trend. Hassan [4] noted forecasting the currency exchange value. Laboissiere et al.
that complex calculations are not particularly effective for [10] developed a model involving correlation analysis and
predicting stock markets. Many trend analysis indicators artificial neural networks (NNs) to predict the stock prices
and prediction methods for financial markets have been of Brazilian electric companies. Lei [11] proposed a wavelet
proposed. Pai [5] used Internet search trends and historical NN prediction method for the stock price trend based on
trading data to predict stock markets using the least squares rough set attribute reduction. Lahmiri [12] used variational
support vector regression model. Lahmiri [6] accurately mode decomposition to forecast the intraday stock price.
2 Mathematical Problems in Engineering

Lahmiri [13] addressed the problem of technical analysis OSCt = DIFt − DEAt ,
information fusion and reported that technical information (1)
fusion in an NN ensemble architecture improves the predic-
tion accuracy. In [14], the authors argued that time series where m = 12, n = 26, and p = 9. The weight number
of stock prices are nonstationary and highly noisy. This led 𝛼 is a fixed value equal to 2/(m + 1). The number of the
the authors to propose the use of a wavelet denoising-based MACD histogram is usually called the MACD bar or OSC.
backpropagation (WDBP) NN for predicting the monthly The analysis process of the cross and deviation strategy of DIF
closing price of the Shanghai composite index. Shynkevich et and DEA includes the following three steps.
al. [15] investigated the impact of varying the input window (i) Calculate the values of DIF and DEA.
length and the highest prediction performance was observed (ii) When DIF and DEA are positive, the MACD line cuts
when the input window length was approximately equal to the signal line in the uptrend, and the divergence is positive,
the forecast horizon. In [16], a prediction model based on there is a buy signal confirmation.
the input/output data plan was developed by means of the (iii) When DIF and DEA are negative, the signal line
adaptive neurofuzzy inference system method representing cuts the MACD line in the downtrend, and the divergence
the fuzzy inference system. Zhou et al. [17] proposed a stock is negative, there is a sell signal confirmation.
market prediction model based on high-frequency data using
generative adversarial nets. Wang et al. [18] used a bimodal 3. MACD-HVIX Weighted by Historical
algorithm with a data-divider to predict the stock index.
In [19], the author used multiresolution analysis techniques Volatility and Its Strategy
to predict the interest rate next-day variation. Using K-line The essence of a good technical indicator is a smooth trading
patterns’ predictive power analysis, Tao et al. [20] found that strategy; i.e., the constructed index must be a stationary
their proposed approach can effectively improve prediction process. We present an empirical study in Section 5. The
accuracy for stock price direction and reduce forecast error. validity and sensitivity of MACD have a strong relationship
We will introduce the concept of moving average con- with the choice of parameters. Different investors choose
vergence divergence (MACD) and help the readers under- different parameters to achieve the best return for different
stand its principle and application in Section 2. Although stocks. In this study, the weight is based on the historical
the MACD oscillator is one of the most popular technical volatility. It is expected that the accuracy and stability of
indicators, it is a lagging indicator. In Section 3, we propose MACD can be improved. The construction formula is as
an improved model called MACD-HVIX to deal with the lag follows:
factor. In Section 4, data for empirical research are described.
Finally, in Section 5, we develop a trading strategy using (EMA − HVIX)m
t (St )
MACD-HVIX and employ actual market data to verify its
∑𝑚
𝑖=1 𝜎𝑡−𝑖 𝜎𝑡
validity and reliability. We also compare the prediction accu- = (EMA − HVIX)m
t−1 + St
racy and cumulative return of the MACD-HVIX histogram ∑𝑚 𝜎
𝑖=0 𝑡−𝑖 ∑ 𝑚
𝑖=0 𝜎𝑡−𝑖
with those of the MACD histogram. The performance of
MACD-HVIX exceeds that of MACD. Therefore, the trading (t > 1) ,
strategy based on the MACD-HVIX index is useful for (EMA − HVIX)m
1 = S1 , (2)
trading. Section 6 presents our conclusion.
(MACD − HVIX)t = (DIF − HVIX)t
2. MACD and Its Strategy = (EMA − HVIX)m n
t (St ) − (EMA − HVIX)t (St ) ,
MACD evolved from the exponential moving average p
(DEA − HVIX)t = EMAt ((DIF − HVIX)t ) ,
(EMA), which was proposed by Gerald Appel in the 1970s. It
is a common indicator in stock analysis. The standard MACD (OSC − HVIX)t = (DIF − HVIX)t − (DEA − HVIX)t .
is the 12-day EMA subtracted by the 26-day EMA, which
is also called the DIF. The MACD histogram, which was Here, the weight changes over time; HVIX is the change index
developed by T. Aspray in 1986, measures the signed distance of the historical volatility of a stock. The HVIX in this paper
between the MACD and its signal line calculated using the is the change index of the volatility in the past days. It is
9-day EMA of the MACD, which is called the DEA. Similar similar to the market volatility index VIX used by the Chicago
to the MACD, the MACD histogram is an oscillator that options exchange. It reflects the panic of the market to a
fluctuates above and below the zero line. The construction certain extent; thus, it is also called the panic index. The above
formula is as follows: process is expressed by the code shown in Algorithm 1.
The analysis process of the cross and deviation strategy
EMAm m
t (St ) = (1 − 𝛼) EMAt−1 + 𝛼 × St (t > 1) , of DIF-HVIX and DEA-HVIX includes the following three
EMAm
1 = S1 , steps.
(i) Calculate the values of DIF-HVIX and DEA-HVIX.
MACDt = DIFt = EMAm n
t (St ) − EMAt (St ) ,
(ii) When DIF-HVIX and DEA-HVIX are positive, the
p
DEAt = EMAt (DIFt ) , MACD-HVIX line cuts the signal line of HVIX in the
Mathematical Problems in Engineering 3

Require: Set up parameters

The stock closing price is 𝑆𝑐𝑙𝑜𝑠𝑒 , the historical volatility index is 𝐻𝑉𝐼𝑋, the length of the closing
stock price data is 𝑁𝐿 , the weight of 𝐸𝑀𝐴 is 𝐴𝑙𝑝ℎ𝑎, the weight of 𝐸𝑀𝐴 𝐻𝑉𝐼𝑋 is 𝐴𝑙𝑝ℎ𝑎 𝐻𝑉𝐼𝑋,
and the time parameters are m and n.
for j=1 to (𝑁𝐿 − 𝑚 + 1) do
sum=0;
for i=j to (j+𝑚 − 2) do
󳨐⇒Generate return series
𝑆
𝑅𝑖 = log ( 𝑐𝑙𝑜𝑠𝑒𝑖+1 )
𝑆𝑐𝑙𝑜𝑠𝑒𝑖
󳨐⇒Sum the return in the past 𝑚 − 1 days
sum=sum+𝑅𝑖
Calculate the mean return in the past 𝑚 − 1 days
sum
R mean=
(𝑚 − 1)
sum=0
for i=j to (j+𝑚 − 2) do
𝑉𝑎𝑟𝑖 = (𝑅𝑖 − 𝑅 𝑚𝑒𝑎𝑛)2
󳨐⇒Sum the variance in the past 𝑚 − 1 days
sum=sum+𝑉𝑎𝑟𝑖 ;
Calculate the standard deviation in the past 𝑚 − 1 days
𝑠𝑢𝑚
𝑆𝑡𝑑𝑗 = √
𝑚−2
for j=1 to (𝑁𝐿 − 2𝑚 + 2) do
sum1=0
for i=j to (j+𝑚 − 1) do
sum1=sum1+𝑆𝑡𝑑𝑖 ;
sum2=0;
for i=j to (j+𝑚 − 2) do
sum2=sum2+𝑆𝑡𝑑𝑖 ;
Calculate 𝐴𝑙𝑝ℎ𝑎 𝐻𝑉𝐼𝑋
𝑠𝑢𝑚2
𝐴𝑙𝑝ℎ𝑎 𝐻𝑉𝐼𝑋𝑗 = 1 −
𝑠𝑢𝑚1
for i=2 to (𝑁𝐿 − 2𝑚 + 2) do
𝐸𝑀𝐴 𝐻𝑉𝐼𝑋𝑖 = (1 − 𝐴𝑙𝑝ℎ𝑎 𝐻𝑉𝐼𝑋𝑖−1 ) × 𝐸𝑀𝐴 𝐻𝑉𝐼𝑋𝑖−1 + 𝐴𝑙𝑝ℎ𝑎 𝐻𝑉𝐼𝑋𝑖−1 × 𝑆𝑐𝑙𝑜𝑠𝑒𝑖+2𝑚−2
Calculate 𝐷𝑖𝑓𝑓 𝐻𝑉𝐼𝑋 and 𝐷𝐸𝐴 𝐻𝑉𝐼𝑋
𝐷𝑖𝑓𝑓 𝐻𝑉𝐼𝑋 = 𝐸𝑀𝐴 𝐻𝑉𝐼𝑋𝑚 − 𝐸𝑀𝐴 𝐻𝑉𝐼𝑋𝑛
Calculate 𝐷𝐸𝐴 𝐻𝑉𝐼𝑋
𝑝
∑ 𝐷𝐼𝐹 𝐻𝑉𝐼𝑋
𝐷𝐸𝐴 𝐻𝑉𝐼𝑋(1) = 𝑖=1
𝑝
2
𝐴𝑙𝑝ℎ𝑎𝑚 =
𝑚+1
for i=2 to length(𝐷𝑖𝑓𝑓 𝐻𝑉𝐼𝑋)−𝑝 + 1 do
𝐷𝐸𝐴 𝐻𝑉𝐼𝑋𝑖 = (1 − 𝐴𝑙𝑝ℎ𝑎𝑚 ) × 𝐷𝐸𝐴 𝐻𝑉𝐼𝑋𝑖−1 + 𝐴𝑙𝑝ℎ𝑎𝑚 × 𝐷𝑖𝑓𝑓 𝐻𝑉𝐼𝑋𝑖+𝑝−1
Calculate OSC HVIX
OSC HVIX = Diff HVIX − DEA HVIX

Algorithm 1: General algorithm for HVIX and EMA-HVIX.

uptrend, and the divergence is positive, there is a buy We use the historical data for the stock “-zgrs-” from
signal confirmation. November 2, 2015, to September 21, 2017, from the Shanghai
stock market. First, we develop the strategy for the new index
(iii) When DIF-HVIX and DEA-HVIX are negative, the and calculate the prediction accuracy and cumulative return
signal line of HVIX cuts the MACD-HVIX line in the of the stock with two different indicators. Then, we compare
downtrend, and the divergence is negative, there is a the accuracy rate and cumulative return. The accuracy here is
sell signal confirmation. calculated according to whether the stock price rises on the
second day. Furthermore, we test a buy-and-hold strategy for
4. Data Description the proposed model. The buy-and-hold strategy is a trading
strategy in which the traders hold the stock for a while instead
We first perform an empirical study on the buy-and-sell of selling it on the next trading day. We use the historical
strategy, which involves buying today and selling tomorrow. data for the stock “-dggf-” from July 27, 2009, to November
4 Mathematical Problems in Engineering

0.05

Candlestick chart
30
0.045
0.04 25
0.035
20
0.03
HVIX

0 50 100 150 200 250 300 350 400 450

0.025 date
12-day EMA
0.02 26-day EMA
0.015 1

MACD histogram
0.01
0
0.005
−1
0
0 50 100 150 200 250 300 350 400 450 500 Sell Buy Buy Buy Sell Buy Buy
date −2
0 50 100 150 200 250 300 350 400 450
HVIX12 date
HVIX26 DIF
DEA
Figure 1: Historical volatility of “-zgrs-” with the buy-and-sell MACD bar
strategy.
Figure 2: Candlestick chart and MACD histogram for “-zgrs-” with
the buy-and-sell strategy. (For interpretation of the references to
color in the figure, the reader is referred to the web version of the
3, 2017, from the Shanghai stock market to test a 5 d buy- article.)
and-hold strategy and use the historical data for the stock “-
payh-” from June 22, 1993, to May 10, 2010, from the Shanghai
stock market to test a 10 d buy-and-hold strategy. The detailed
divergence is positive. We sell the stock when the DEA cuts
trading strategy is similar to the buy-and-sell strategy. Here,
the DIF in a downtrend, and the divergence is negative. As
we use m = 12, n = 26, and p = 9.
shown in Figure 2, we sell the stock on days 155 and 355 and
buy the stock on days 212, 290, 310, 381, and 393. The buy-and-
5. Empirical Results sell signals in the candlestick chart and the MACD histogram
are shown in Figure 3.
5.1. Empirical Results of Buy-and-Sell Strategy. From the “- Figure 4 shows the candlestick chart and MACD his-
zgrs-” stock data chosen in Section 4, we calculate the HVIX togram of HVIX. In the candlestick chart, the blue line
of the past m days and the past n days. A higher stock represents the 12-d EMA-HVIX, and the red line represents
index means that investors feel anxiety regarding the stock the 26-d EMA-HVIX. In the MACD-HVIX histogram, the
market, and a lower stock index means that the rate of change solid line represents the DIF-HVIX, the dotted line represents
of the stock price will decrease. Figure 1 shows the HVIX the DEA-HVIX, and the histogram represents the MACD-
index. HVIX bar. According to the strategy described in Section 3,
Next, using the calculated volatility index, we calculate we buy the stock when the DIF-HVIX and DEA-HVIX are
the weight of the EMA formula in Section 3 and obtain the positive, the DIF-HVIX cuts the DEA-HVIX in an uptrend,
values of MACD-HVIX, DEA-HVIX, and OSC. and the divergence is positive, and we sell the stock when
Figure 2 shows the candlestick chart and MACD his- the DEA-HVIX cuts the DIF-HVIX in a downtrend, and the
togram. In the candlestick chart, the blue line represents divergence is negative. As shown in Figure 4, we sell the
the 12-d EMA, and the red line represents the 26-d EMA. stock on days 118 and 187 and buy the stock on days 222,
Candlesticks are usually composed of a red and green body, 231, 241, 243, 292, 415, and 447. The buy-and-sell signals in
as well as an upper wick and a lower wick. The area between the candlestick chart and the MACD histogram are shown in
the opening and the closing prices is called the body, and Figure 5.
price excursions above or below the real body are called the To verify the stability of the new indicator, we compare
wick. The body indicates the opening and closing prices, and the MACD and MACD-HVIX in Figure 6. The MACD and
the wick indicates the highest and lowest traded prices of a MACD-HVIX have basically the same trend and the stability
stock during the time interval represented. For a red body, of the MACD-HVIX is better than that of the MACD.
the opening price is at the bottom, and the closing price is at Table 1 shows a comparison of the specific values of
the top. For a green body, the opening price is at the top, and the buying-selling points for the MACD index and MACD-
the closing price is at the bottom. HVIX index, as well as a comparison of the predicted and
In the MACD histogram, the solid line represents the DIF, actual trends. Here, we see that the prediction accuracy of
the dotted line represents the DEA, and the histogram rep- MACD-HVIX is 0.667 and that of MACD is 0.4286. By
resents the MACD bar. According to the strategy described using the proposed indicator, we can improve the prediction
in Section 3, we buy the stock when the DIF and DEA accuracy by 55.55% compared with the traditional MACD.
are positive, the DIF cuts the DEA in an uptrend, and the The “-Price-” in the table represents the closing price of
Mathematical Problems in Engineering 5

23 23

Candlestick chart

Candlestick chart
22 22
21 21
20 20
19 19
140 145 150 155 160 165 170 200 205 210 215 220 225
date date
12-day EMA 26-day EMA 12-day EMA 26-day EMA

1 1
MACD histogram

MACD histogram
0 0

−1 −1
Sell Buy
−2 −2
140 145 150 155 160 165 170 200 205 210 212 215 220 225
date date
DIF DEA MACD bar DIF DEA MACD bar
28 28
Candlestick chart

Candlestick chart
26 26

24 24

22 22
275 280 285 290 295 300 305 295 300 305 310 315 320 325
date date
12-day EMA 26-day EMA 12-day EMA 26-day EMA

1
MACD histogram

1
MACD histogram

0 0

−1 −1
Buy Buy
−2 −2
275 280 285 290 295 300 305 295 300 305 310 315 320 325
date date
DIF DEA MACD bar DIF DEA MACD bar
28 30
Candlestick chart

Candlestick chart

26 28

26
24
24
22
340 345 350 355 360 365 370 370 375 380 385 390 395
date date
12-day EMA 26-day EMA 12-day EMA 26-day EMA

1
MACD histogram

1
MACD histogram

0 0

−1 −1
Sell Buy
−2 −2
340 345 350 355 360 365 370 370 375 380 381 385 390 395
date date

DIF DEA MACD bar DIF DEA MACD bar

30
Candlestick chart

28

26

24

380 385 390 395 400 405

date
12-day EMA 26-day EMA
1
MACD histogram

−1
Buy
−2
380 385 390 393 395 400 405
date
DIF DEA MACD bar

Figure 3: Buy-and-sell signals in the candlestick chart and MACD histogram for the buy-and-sell strategy.
6 Mathematical Problems in Engineering

Table 1: Comparison of the specific values of the buying-selling points for the buy-and-sell strategy.

Test date Price (Test date) Price (Test date + 1 d) Predicted trend Actual trend
155 20.39 20.42 ↓ ↑
212 21.82 21.93 ↑ ↑
290 25.32 25.38 ↑ ↑
MACD 310 26.44 26.50 ↑ ↑
355 24.51 24.62 ↓ ↑
381 27.93 27.45 ↑ ↓
393 29.00 28.48 ↑ ↓
Win rate 0.4286
118 22.24 21.90 ↓ ↓
187 20.80 20.66 ↓ ↓
222 21.90 21.84 ↑ ↓
231 21.80 21.67 ↑ ↓
MACD-HVIX 241 21.65 21.58 ↑ ↓
243 21.78 21.83 ↑ ↑
292 25.18 25.58 ↑ ↑
415 27.98 28.80 ↑ ↑
447 29.27 29.51 ↑ ↑
Win rate 0.6667

5.2. Empirical Results of Buy-and-Hold Strategy. Using the “-

Candlestick chart

30
dggf-” stock data chosen in Section 4, we first investigate the
25
buy or sell points for both the indicators with the buy-and-
hold strategy applied for 5 d. Then, we compare the prediction
20
accuracy between the two indicators. The MACD histogram
shown in Figure 7 indicates the buy-and-sell points; we
0 50 100 150 200 250 300 350 400 450
date
should buy the stock at a buy point on days 391, 1,071, 1,181,
12-day EMA-HVIX 1,326, and 1,481, and sell the stock at a sell point on days 791,
26-day EMA-HVIX 881, and 911. The prediction situation is shown in Table 2.
The MACD-HVIX histogram in Figure 8 indicates the
MACD-HVIX histogram

1 buy-and-sell points. We should buy the stock at a buy point

on days 371, 1,201, 1,331, and 1,561 and sell the stock at a sell
0 point on days 751 and 771. The prediction situation is shown
−1 Buy in Table 2. A comparison between MACD and MACD-HVIX
Buy is shown in Figure 9.
−2 Sell Sell Buy Buy Buy Buy Buy Table 2 shows a comparison of the specific values of
0 50 100 150 200 250 300 350 400 450 the buying-selling points for the MACD and MACD-HVIX
date indices with the buy-and-hold strategy for 5 d, as well as
DIF-HVIX a comparison of the predicted and actual trends. Here, we
DEA-HVIX observe that the prediction accuracy of MACD-HVIX is
MACD-HVIX bar 0.8333 and that of MACD is 0.6250. By using the proposed
Figure 4: Candlestick chart and MACD-HVIX histogram for “-
indicator, we can improve the prediction accuracy by 33.33%
zgrs-” with the buy-and-sell strategy. (For interpretation of the compared with the traditional MACD. The “-Price-” in the
references to color in this figure, the reader is referred to the web table represents the closing price of the stock.
version of the article.) Next, using the “-payh-” stock data chosen in Section 4,
we investigate the buy or sell points for both the indicators
with the buy-and-hold strategy applied for 10d. Then, we
compare the prediction accuracy between the two indicators.
stock. Next, we compare the cumulative returns for the two The MACD histogram in Figure 10 indicates the buy-and-sell
indicators. According to the trading points shown in the table, points. We should buy the stock at a buy point on day 901,
we perform a simulation test. We assume that the initial fund and sell the stock at a sell point on days 621, 1,971, 2,071, 2,291,
is 1 million. The cumulative returns under the two indexes 2,431, and 2,661. The prediction situation is shown in Table 3.
are 1.1136 million and 1.3365 million, for MACD and MACD- The MACD-HVIX histogram in Figure 11 indicates the
HVIX indices, respectively. buy-and-sell points. We should buy the stock at a buy point
Mathematical Problems in Engineering 7

Candlestick chart 24 24

Candlestick chart
22 22

20 20

18 18

105 110 115 120 125 130 175 180 185 190 195 200
date date
12-day EMA-HVIX 12-day EMA-HVIX
26-day EMA-HVIX 26-day EMA-HVIX

1
MACD-HVIX histogram

MACD-HVIX histogram
1

0
0

−1
−1

−2 Sell
−2 Sell
175 180 185 187 190 195 200
105 110 115 118 120 125 130
date date
DIF-HVIX
DIF-HVIX
DEA-HVIX
DEA-HVIX
MACD-HVIX bar
MACD-HVIX bar

24
24 Candlestick chart
Candlestick chart

22
22
20
20
18
18
220 225 230 235 240 245
210 215 220 225 230 235 date
date
12-day EMA-HVIX
12-day EMA-HVIX
26-day EMA-HVIX
26-day EMA-HVIX
1
MACD-HVIX histogram
MACD-HVIX histogram

1
0
0
−1
−1
−2 Buy
−2 Buy
220 225 230 231 235 240 245
210 215 220 222 225 230 235 date
date

DIF-HVIX
DIF-HVIX
DEA-HVIX
DEA-HVIX
MACD-HVIX bar
MACD-HVIX bar

24 24
Candlestick chart

Candlestick chart

22 22

20 20

18 18

230 235 240 245 250 255 230 235 240 245 250 255
date date

12-day EMA-HVIX 12-day EMA-HVIX

26-day EMA-HVIX 26-day EMA-HVIX
MACD-HVIX histogram

MACD-HVIX histogram

1 1

0 0

−1 −1

−2 Buy Buy
−2
230 235 240 241 245 250 255 230 235 240 243 245 250 255
date date

DIF-HVIX DIF-HVIX
DEA-HVIX DEA-HVIX
MACD-HVIX bar MACD-HVIX bar

Figure 5: Continued.
8 Mathematical Problems in Engineering

28
30

Candlestick chart
Candlestick chart
26
28
24 26
22 24

20 22
280 285 290 295 300 305 400 405 410 415 420 425 430
date
date
12-day EMA-HVIX
12-day EMA-HVIX
26-day EMA-HVIX
26-day EMA-HVIX

MACD-HVIX histogram
MACD-HVIX histogram

1 1

0 0

−1 −1

Buy −2 Buy
−2
280 285 290 292 295 300 305 400 405 410 415 420 425 430
date date
DIF-HVIX DIF-HVIX
DEA-HVIX DEA-HVIX
MACD-HVIX bar MACD-HVIX bar

30
Candlestick chart

28

26

24

22
435 440 445 450 455 460
date
12-day EMA-HVIX
26-day EMA-HVIX
MACD-HVIX histogram

−1

−2 Buy

435 440 445 447 450 455 460

date

DIF-HVIX
DEA-HVIX
MACD-HVIX bar

Figure 5: Buy-and-sell signals in the candlestick chart and the MACD-HVIX histogram for the buy-and-sell strategy.

on day 901 and sell the stock at a sell point on days 2,071, 2,421, has a length of n are 𝑂(𝑁) and 𝑂(𝑁2 ), respectively. In
2,661, and 2,741. The prediction situation is shown in Table 3. terms of trend prediction processing time, the average time
A comparison of MACD and MACD-HVIX is presented in required to process a buy-and-sell strategy, a buy-and-hold
Figure 12. strategy for 5 days, and a buy-and-hold strategy for 10 days
Table 3 shows the comparison of the specific values of with the MACD approach (MACD-HVIX) are, respectively,
the buying-selling points for the MACD and MACD-HVIX 1.25 (1.51), 1.12 (1.35), and 1.41 seconds (1.58) using Matlab
indices with the buy-and-hold strategy applied for 5 d, as well R2017b on an Intel(R) Core(TM) i5-6200 CPU @ 2.30GHz
as a comparison of the predicted and actual trends. Here, processor.
we observe that the prediction accuracy of MACD-HVIX
is 0.8 and that of MACD is 0.7143. By using the proposed
6. Conclusion
indicator, we can improve the prediction accuracy by 12%
compared with the traditional MACD. The “-Price-” in the As indicated by Tables 1, 2, and 3, we buy-and-sell stock based
table represents the closing price of the stock. on improved MACD; then we found all the accuracy is higher
than that before the improvement. Therefore, the improved
5.3. Computational Complexity. The computational complex- model has higher maneuverability in securities investment
ity of the MACD and MACD-HVIX for a stock which and allows investors to capture every buy-and-sell points in
Mathematical Problems in Engineering 9

Table 2: Comparison of the specific values of the buying-selling points with the buy-and-hold strategy applied for 5 d.

Test date Price (Test date) Price (Test date + 5 d) Predicted trend Actual trend
391 13.04 15.70 ↑ ↑
791 4.81 5.19 ↓ ↑
881 5.55 5.61 ↓ ↑
911 5.71 5.59 ↓ ↓
MACD
1071 6.57 6.74 ↑ ↑
1181 6.66 6.95 ↑ ↑
1326 10.70 11.93 ↑ ↑
1481 17.28 13.51 ↑ ↓
Win rate 0.6250
371 11.68 12.00 ↑ ↑
751 5.53 5.75 ↓ ↑
771 5.76 5.62 ↓ ↓
MACD-HVIX
1201 7.66 7.83 ↑ ↑
1331 11.93 12.23 ↑ ↑
1561 16.89 19.05 ↑ ↑
Win rate 0.8333

0.6
Candlestick chart

20
MACD bar and MACD-HVIX bar

0.4
10
0.2
0
0 200 400 600 800 1000 1200 1400 1600 1800
0 date
12-day EMA
26-day EMA
−0.2 4
MACD histogram

−0.4 2

0
−0.6
0 50 100 150 200 250 300 350 400 450 500 Sell
Buy Sell Sell Buy Buy Buy Buy
−2
date 0 200 400 600 800 1000 1200 1400 1600 1800
MACD bar date
DIF
MACD-HVIX bar DEA
MACD bar
Figure 6: Comparison of the two indicators for the buy-and-sell
strategy. Figure 7: Candlestick chart and MACD histogram for “-dggf-” with
the buy-and-hold strategy applied for 5 d. (For interpretation of the
references to color in this figure, the reader is referred to the web
version of the article.)
the market. For both the buy-and-sell strategy and the buy-
and-hold strategy, the empirical results indicated that the
proposed model can make more precise predictions than
the traditional model. The proposed model was tested with
three different stocks and it generated the high prediction investors. In future research, we will investigate other factors
accuracy for all the cases. In addition, while a smoothing for the model by constantly updating the data and the training
index is used to construct the MACD index and the impact of model to obtain a better prediction effect.
the past price declines exponentially, the MACD-HVIX does
not have this property. Although the MACD-HVIX index is
improved compared with the MACD index, the stationarity Data Availability
of the MACD-HVIX index is difficult prove theoretically.
Test shows that it is stable; however, in the ever-changing The data used to support the findings of this study are
market, an abnormal situation can cause incalculable losses to available from the corresponding author upon request.
10 Mathematical Problems in Engineering

Table 3: Comparison of the specific values of the buying-selling points with the buy-and-hold strategy applied for 10 d.

Test date Price (Test date) Price (Test date + 10 d) Predicted trend Actual trend
621 6.59 6.23 ↓ ↓
901 18.87 19.05 ↑ ↑
1971 13.05 13.22 ↓ ↑
MACD 2071 12.25 10.31 ↓ ↓
2291 11.11 10.52 ↓ ↓
2431 10.56 10.81 ↓ ↑
2661 8.51 8.23 ↓ ↓
Win rate 0.7143
901 18.87 19.05 ↑ ↑
2071 12.25 10.31 ↓ ↓
MACD-HVIX 2421 11.30 10.56 ↓ ↓
2661 8.51 8.23 ↓ ↓
2741 6.95 7.19 ↓ ↑
Win rate 0.8000
Candlestick chart

20

10

0
0 200 400 600800 1000 1200 1400 1600 1800
date
12-day EMA-HVIX
26-day EMA-HVIX
MACD-HVIX histogram

0
Buy Sell Sell Buy Buy Buy
−2
0 200 400 600 800 1000 1200 1400 1600 1800
date
DIF-HVIX
DEA-HVIX
MACD-HVIX bar

Figure 8: Candlestick chart and MACD-HVIX histogram for “-dggf-” with the buy-and-hold strategy for 5 d. (For interpretation of the
references to color in this figure, the reader is referred to the web version of the article.)

1.5
MACD bar and MACD-HVIX bar

0.5

−0.5

−1

−1.5
0 200 400 600 800 1000 1200 1400 1600 1800 2000
date
MACD bar
MACD-HVIX bar

Figure 9: Comparison of the two indicators with the buy-and-hold strategy applied for 5 d.
Mathematical Problems in Engineering 11

3
Candlestick chart

40

MACD bar and MACD-HVIX bar

2
20
1
0
0 500 1000 1500 2000 2500 3000 3500 4000
date 0
12-day EMA
26-day EMA
10 −1
MACD histogram

5
−2
0
−3
Se ll Buy Sell Sell Sell Sell Sell 0 500 1000 1500 2000 2500 3000 3500 4000 4500
−5
0 500 1000 1500 2000 2500 3000 3500 4000 date
date MACD bar
DIF MACD-HVIX bar
DEA
MACD bar Figure 12: Comparison of the two indicators with the buy-and-hold
Figure 10: Candlestick chart and MACD histogram for “-payh-” strategy applied for 10 d.
with the buy-and-hold strategy applied for 10 d. (For interpretation
of the references to color in this figure, the reader is referred to the
web version of the article.) author (J.S. Kim) expresses thanks for the support from the
BK21 PLUS program.
Candlestick chart

40 References
[1] F. Schmitt, D. Schertzer, and S. Lovejoy, “Multifractal Fluc-
20
tuations in Finance,” International Journal of Theoretical and
Applied Finance, vol. 03, no. 03, pp. 361–364, 2000.
0
0 500 1000 1500 2000 2500 3000 3500 4000 [2] T. J. Moskowitz, Y. H. Ooi, and L. H. Pedersen, “Time series
date momentum,” Journal of Financial Economics, vol. 104, no. 2, pp.
12-day EMA-HVIX 228–250, 2012.
26-day EMA-HVIX
MACD-HVIX histogram

[3] C. S. Asness, T. J. Moskowitz, and L. H. Pedersen, “Value and

10
Momentum Everywhere,” Journal of Finance, vol. 68, no. 3, pp.
5 929–985, 2013.
[4] M. R. Hassan, “A combination of hidden Markov model and
0 fuzzy model for stock market forecasting,” Neurocomputing, vol.
Buy Sell Sell Sell Sell
72, no. 16-18, pp. 3439–3446, 2009.
−5 [5] P. Pai, L. Hong, and K. Lin, “Using Internet Search Trends
0 500 1000 1500 2000 2500 3000 3500 4000
date and Historical Trading Data for Predicting Stock Markets by
DIF-HVIX the Least Squares Support Vector Regression Model,” Com-
DEA-HVIX putational Intelligence and Neuroscience, vol. 2018, Article ID
MACD-HVIX bar 6305246, 15 pages, 2018.
Figure 11: Candlestick chart and MACD-HVIX histogram for [6] S. Lahmiri, “Minute-ahead stock price forecasting based on
“-payh-” with the buy-and-hold strategy applied for 10 d. (For singular spectrum analysis and support vector regression,”
interpretation of the references to color in this figure, the reader is Applied Mathematics and Computation, vol. 320, pp. 444–451,
referred to the web version of the article.) 2018.
[7] P. Singh and B. Borah, “Forecasting stock index price based on
M-factors fuzzy time series and particle swarm optimization,”
International Journal of Approximate Reasoning, vol. 55, no. 3,
Conflicts of Interest pp. 812–833, 2014.
The authors declare that there are no conflicts of interest [8] S. Lahmiri and M. Boukadoum, “Intelligent Ensemble Forecast-
regarding the publication of this paper. ing System of Stock Market Fluctuations Based on Symetric and
Asymetric Wavelet Functions,” Fluctuation and Noise Letters,
vol. 14, no. 4, 2015.
Acknowledgments [9] S. R. Das, D. Mishra, and M. Rout, “A hybridized ELM
using self-adaptive multi-population-based Jaya algorithm for
The first author (Jian Wang) was supported by the China currency exchange prediction: an empirical assessment,” Neural
Scholarship Council (201808260026). The corresponding Computing and Applications, pp. 1–24, 2018.
12 Mathematical Problems in Engineering

[10] L. A. Laboissiere, R. A. S. Fernandes, and G. G. Lage, “Maximum

and minimum stock price forecasting of Brazilian power distri-
bution companies based on artificial neural networks,” Applied
Soft Computing, vol. 35, pp. 66–74, 2015.
[11] L. Lei, “Wavelet Neural Network Prediction Method of Stock
Price Trend Based on Rough Set Attribute Reduction,” Applied
Soft Computing, vol. 62, pp. 923–932, 2018.
[12] S. Lahmiri, “Intraday stock price forecasting based on varia-
tional mode decomposition,” Journal of Computational Science,
vol. 12, pp. 23–27, 2016.
[13] S. Lahmiri, “A Technical Analysis Information Fusion Approach
for Stock Price Analysis and Modeling,” Fluctuation and Noise
Letters, vol. 17, no. 01, p. 1850007, 2018.
[14] J.-Z. Wang, J.-J. Wang, Z.-G. Zhang, and S.-P. Guo, “Forecasting
stock indices with back propagation neural network,” Expert
Systems with Applications, vol. 38, no. 11, pp. 14346–14355, 2011.
[15] Y. Shynkevich, T. M. McGinnity, S. A. Coleman, A. Belatreche,
and Y. Li, “Forecasting price movements using technical indica-
tors: Investigating the impact of varying input window length,”
Neurocomputing, vol. 264, pp. 71–88, 2017.
[16] I. Svalina, V. Galzina, R. Lujić, and G. Šimunović, “An adaptive
network-based fuzzy inference system (ANFIS) for the fore-
casting: the case of close price indices,” Expert Systems with
Applications, vol. 40, no. 15, pp. 6055–6063, 2013.
[17] X. Zhou, Z. Pan, G. Hu, S. Tang, and C. Zhao, “Stock Market Pre-
diction on High-Frequency Data Using Generative Adversarial
Nets,” Mathematical Problems in Engineering, vol. 2018, Article
ID 4907423, 11 pages, 2018.
[18] Z. Wang, J. Hu, and Y. Wu, “A Bimodel Algorithm with Data-
Divider to Predict Stock Index,” Mathematical Problems in
Engineering, vol. 2018, Article ID 3967525, 14 pages, 2018.
[19] S. Lahmiri, “Interest rate next-day variation prediction based
on hybrid feedforward neural network, particle swarm opti-
mization, and multiresolution techniques,” Physica A: Statistical
Mechanics and its Applications, vol. 444, pp. 388–396, 2016.
[20] L. Tao, Y. Hao, H. Yijie, and S. Chunfeng, “K-Line Patterns’
Predictive Power Analysis Using the Methods of Similarity
Match and Clustering,” Mathematical Problems in Engineering,
vol. 2017, Article ID 3096917, 11 pages, 2017.
Advances in Advances in Journal of The Scientific Journal of
Operations Research
Hindawi
Decision Sciences
Hindawi
Applied Mathematics
Hindawi
World Journal
Hindawi Publishing Corporation
Probability and Statistics
Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 http://www.hindawi.com
www.hindawi.com Volume 2018
2013 www.hindawi.com Volume 2018

International
Journal of
Mathematics and
Mathematical
Sciences

Journal of

Hindawi
Optimization
Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Submit your manuscripts at

www.hindawi.com

International Journal of
Engineering International Journal of
Mathematics
Hindawi
Analysis
Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Journal of Advances in Mathematical Problems International Journal of Discrete Dynamics in

Complex Analysis
Hindawi
Numerical Analysis
Hindawi
in Engineering
Hindawi
Differential Equations
Hindawi
Nature and Society
Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

International Journal of Journal of Journal of Abstract and Advances in

Stochastic Analysis
Hindawi
Mathematics
Hindawi
Function Spaces
Hindawi
Applied Analysis
Hindawi
Mathematical Physics
Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018