CN112257770B

CN112257770B - K-line form recognition method

Info

Publication number: CN112257770B
Application number: CN202011119572.3A
Authority: CN
Inventors: 廖钧
Original assignee: Sichuan Changhong Electric Co Ltd
Current assignee: Sichuan Changhong Electric Co Ltd
Priority date: 2020-10-19
Filing date: 2020-10-19
Publication date: 2022-05-06
Anticipated expiration: 2040-10-19
Also published as: CN112257770A

Abstract

The invention relates to the technical field of computer software, and discloses a K-line shape recognition method which is used for solving the problem of K-line shape recognition. The invention comprises the following steps: (1) accessing a financial market data source; (2) finding out the high and low points of the closing price in the K line stage by using a ZIG function; (3) connecting high and low points of adjacent stages, and performing K line shape classification according to the relation of the high and low points of the K line stage to form a K line shape library; (4) setting the K-line state to be found, traversing the K-line state library, and finding out the corresponding K-line state in the K-line state library. The invention is suitable for K-line shape recognition.

Description

K-line form recognition method

Technical Field

The invention relates to the technical field of computer software, in particular to a K-line form identification method.

Background

The K-line chart is in the RibendeChuan curtain age, is used by traders in Japan rice market at that time to record market conditions and price fluctuations of rice market, and is introduced into financial markets due to the exquisite unique marking mode. The drawing method of the K line graph in the financial market comprises four data, namely the opening price, the highest price, the lowest price and the closing price, and all the K lines are unfolded around the four data to reflect the situation and the price information of the great situation.

Through the K line graph, the market condition performance of each day or a certain period can be completely recorded, after the price passes through a period of time, a special area or form is formed on the graph, and different forms show different meanings. We can feel some regularity from the change of these forms. Since the expression of the K-line graph is intuitive and easy to understand, the K-line graph analysis is also an important tool and means for technical analysis.

In the past analysts manually analyzed K-line graphs. The main basis of the method is the Eisenet wave theory. The theory holds that the K-line plot repeats a pattern continuously, each cycle consisting of 5 rising waves and 3 falling waves. Many of the eigter wave theories suffer from a problem, is it done by one wave and started by another? Sometimes the first wave is seen and the second wave is seen. Poor in millili and lost in thousand miles. The consequences of a misreading can be quite severe. Even if it is a complete wave, there is no clear definition, and the number of rise and fall in the actual K-line graph mostly does not appear in the mechanical mode of five-rise and three-fall. But wave theorists misinterpret that some of the rises and falls should not be counted into the waves. Waves are completely subjective at will.

Because the wave theory of the Aidit is a set of subjective analysis tools, no objective criteria exist. The problems of high error rate of K-line state recognition, low efficiency and the like are easily caused by the reasons of inconsistent levels of analysts, large number of K-lines and the like. People have therefore begun to think about machines to assist humans in recognizing K-line morphology. However, the problem encountered at present is that the K-ray is non-standard due to its complex and various forms. Wave theory cannot give the machine a quantifiable recognition criterion, so that the key points of the stage cannot be identified: the highest point and the lowest point are not ideal even if machine learning and artificial intelligence are applied.

Disclosure of Invention

The technical problem to be solved by the invention is as follows: a K-line shape recognition method is provided to solve the problem of K-line shape recognition.

In order to solve the problems, the invention adopts the technical scheme that: a K-line shape recognition method, comprising:

(1) accessing a financial market data source;

(2) finding out the high and low points of the closing price in the K line stage by using a ZIG function;

(3) connecting high and low points of adjacent stages, and then classifying K line forms according to the relation of the high and low points of the K line stage to form a K line form library;

(4) setting the K-line state to be found, traversing the K-line state library, and finding out the corresponding K-line state in the K-line state library.

Specifically, in the step (2), the method for finding the high and low points of the phase closing price by using the ZIG function is as follows,

ZIG function

Zig-zag direction, usage:

the function expression: ZIG (K, N, ABS);

the expression means:

turning to when the price change exceeds N%,

the meaning of K is:

0, opening price, 1, highest price, 2, lowest price, 3, closing price, 4, adopting lowest price at low point and adopting highest price at high point.

If ABS is 0 or omitted, relative ZIG steering is indicated, otherwise absolute ZIG steering is indicated.

For example:

ZIG (3, 5); the meaning of the expression is: ZIG turn 5% of closing price;

ZIG (3,0.5, 1); the meaning of the expression is: 0.5-membered absolute ZIG turn to closing price.

Take ZIG (3,5) as an example, 3 represents a closing price, and 5 represents 5%.

Further, after the stage high and low points are found in the step (2), the high and low points are respectively defined as a next-new high point, a next-new low point, a new high point and a new low point; the method for connecting the high points and the low points of the adjacent stages in the step (3) comprises the following steps: connecting a high point with a high point, and connecting a low point with a low point; and when the K line shape is classified, judging the sizes of the front and rear high and low points to realize the shape classification.

Further, the K-line shape is divided into four first-order shapes of greater than, less than, upward and downward in step (3), and if the new high point, the next-new low point, the new high point and the new low point are respectively recorded as A, B, C, D, the greater shape is expressed as: a > -C, B < ═ D; the less than class morphology is represented as: a < C, B > D; the upward morphology is represented as: a < C, B < ═ D; the downward morphology is represented as: a > -C, B > D.

The invention has the following beneficial effects: the invention introduces a brand-new K-line shape classification method when identifying the K-line shape, which is completely different from the prior method of realizing shape classification by using a calculation formula of the slopes of adjacent key points in the K-line or realizing shape classification by using convolution comparison of each position point in the K-line. Compared with other technologies, the method for finding the stage high and low points by using the technology can be completely quantized, and the technical parameters can be adjusted as required. The K-line morphology classification covers all morphologies in all and the classification can be extended as needed.

Drawings

FIG. 1 is a flowchart illustrating K-ray shape recognition according to an embodiment;

FIG. 2 is a schematic view of a larger scale than the figure;

FIG. 3 is a schematic diagram of a smaller than configuration;

FIG. 4 is a schematic view of the upward configuration;

FIG. 5 is a schematic view of the downward configuration.

Detailed Description

The invention finds out the high and low points of the stage by using a ZIG function through a computer and a steering principle when the relative variation of the price exceeds N percent. The problem that the highest point and the lowest point of the phase cannot be accurately found due to the fact that no quantitative standard exists in K-line state recognition is solved. After finding out the highest point and the lowest point, connecting the front and the back two adjacent highest points and connecting the front and the back two adjacent lowest points, the form of the K-line graph can be divided into one order: greater than, less than, upward, downward; the classification of morphology covers all K-line morphologies and can continue to be extended. Only one more adjacent highest point and one more lowest point are respectively added on the basis of the first-order form to form a second-order form: greater than form, greater than less than form, greater than upward form, greater than downward form, less than upward form, less than downward form, greater than form upward, less than form upward, upward form downward, greater than form downward, less than form downward, upward form downward, downward form downward. The second-order forms are 16 types, and are subdivided on the basis of the first-order forms. In the future, when a certain specified form appears again, the specified form can be quickly identified by a traversal method. Compared with manual work, the system improves efficiency and accuracy. Compared with other technologies, the method for finding the stage high and low points by using the technology can be completely quantized, and the technical parameters can be adjusted as required. The K-line morphology classification covers all morphologies in all and the classification can be extended as needed.

The embodiment provides a K-ray shape recognition method, as shown in fig. 1, which specifically includes the following steps:

(1) accessing a financial market data source;

i. if the following sequence is coiled, 100,96,104,105,99,107.

First, find the first number of deviations from 100 of 5%, obviously the first is 105 (point a), then 105 (point a), is greater than 100, then determine the first ZIG turn as a turn, at which time the apex of Λ should be found, first determine 105 if it is this ZIG break, as long as the first following point greater than 105 and all points within the interval of 105 this point, then this point is less than (1-0.05) × 105, then 105 is the ZIG break, if not, then see if the first point greater than 105 (assuming it is point B) is the ZIG break, the determination is the same as above, first following point greater than B is found, then this point is set as point C, then whether the minimum between B, C is less than 0.95B, if it is then B is the minimum, then either C is the C point or not.

The upper closing price is 99 met, so the first ZIG break point is 105,

further examples are 100,96,104,105,103,102,101,109,103,110,105.

Then this ZIG break point is 109.

But sometimes such situations are encountered, as also described above,

100,96,104,105,103,102,101,109,105,110,109；

thus, there is no zig (3,5) that corresponds to the line graph.

At this time, we can regard the last maximum (or minimum) as a temporary break point, which may not be the case as the curve lengthens as the subsequent values increase.

Therefore, the last break point of the ZIG may not be fixed at all times.

The formula: ZIG (3, PER)

ZIG outputs a numerical value, not a straight line. Only the inflection point has output, and the rest points are smooth connecting lines between the two inflection points. It makes no sense to say that the region is not an inflection point. The ZIG function expresses the inflection point.

Trajectory of ZIG function

The definition of the ZIG function is: the ZIG (K, N) function turns when the value K varies by more than N%.

The trajectory of this function is now known, taking the sun ZIG (3,5) as an example.

(a1) When the closing price rises and falls beyond 5%, the highest or lowest closing price is generated, and the steering occurs, the trajectory of the ZIG function is a straight line sequentially connecting each high point and each low point. The ZIG values for the days in the middle of the high and low points are linear interpolation values.

(a2) The last trading day does not reach 5% of the last highest or lowest closing price point, and the function track of the period is divided into two cases (taking the latest lowest closing price valley point as an example):

1) this period rises all the time without falling back and dropping the latest minimum closing price. The closing price of the last trading day is the ZIG function value of the day, the ZIG track of the period is that the closing price of the day and the lowest closing price draw a straight line, and the ZIG value of each day in the middle is the linear interpolation value of the closing price.

2) The latest minimum closing price is broken by falling back and falling after the period of rising. The ZIG function trajectory was the same as above for the days before the nearest valley point fell again and turned back up. When the steering return rise occurs again, the closing price of the dish is the lowest point in the day before the return rise, the closing price becomes a new steering valley point, and the original valley point disappears. The ZIG function trajectory is a connecting straight line from the newly turned valley point to the front high point. Therefore, after the lowest closing price before the falling and the breaking occurs, before the rising and the turning do not occur, the new low price is not taken as the turning point, and the original valley point becomes the turning point. This turning point disappears only if a turn occurs.

The change in function value of ZIG (3,15) is illustrated below with two examples in 2010.

(1)000989 JIUZHITANG.

The vertex 15.77 yuan appears at 22 month 2010. And 5,10, closing the tray by 13.28 yuan, and changing the ZIG to 13.28 yuan, wherein 15.79% of the top 15.77 yuan is dropped on the same day, and the condition of 15% of the top is met. And 5, 12 disc closing is carried out again and the disc falls to 12.34 yuan, the ZIG is equal to 12.34, the ZIG line is a straight line between 12.34 and a vertex 15.77 on the day, and the ZIG on the middle day is a linear interpolation value of the two points.

After 5 months and 12 days, the rice will become a valley point after one-time rising and 5 months and 12 days. And 6,10, closing 13.38, and keeping the ZIG equal to 13.38, wherein the ZIG line is a straight line between 13.38 of the day and 12.34 of 5 months and 12 days, and the ZIG of the middle day is a linear interpolation value of the two points.

And 6,10, then the product falls back, and 6, 29, the product takes 12.09 and falls through 12.34 of 5, 12. Month 7 and 5 reached a minimum of 10.98. During this period the day ZIG line was 12.34 for 5 months 12 to the day's inventory. It is noted here that after month 6 29, month 5 12 was not a valley point, but the system did not shift the valley point to a new low point, but rather a turning point remained at month 5 12.

6-rise in 7 months, closing the tray by 11.17, increasing the stock price by 11.17, forming a new valley point in 7 months 5, forming a ZIG line from 7 months 5 to 4 months 22, and leaving no turning point in 5 months 12.

From the above records, it can be explained that the lowest point of the ZIG function varies day by day. If today is lower than yesterday, today is the lowest point of the function. If today rises back, a valley point occurs today. Although the valley point appears, after falling back and passing the valley point in the later period, the former valley point becomes the turning point first, and finally the valley point disappears, and the lower point starting to rise back is taken as the valley point. If you buy on day 5, 12, then the loss will be lost by day 29, 6.

(2)002319 Letongguan. And 20 months later, the valley point appears in the period of 5 months, the stock price is reduced in the period of 6 months, 29 months later, the width of the valley point is increased by 16.67, and the turning condition is achieved. Month 6 30 falls back to the tray 21.04 so that month 6 29 becomes a vertex. Later days, the temperature gradually rose back and broke 21.72 in month 29, and the closing of 24.15 in month 2 in month 7, but ZIG was not from 18.60 to 24.15, but was still 18.60 to 21.72, and there was a turning point at 21.72, and then to 24.15. Month 7 5 falls back to tray 23.53, at which time 24.15 of month 7 and 2 become the apex.

(3) After high and low points of the phase are found out, the high and low points are respectively defined as a next new high point A, a next new low point B, a new high point C and a new low point D, the high point is connected with the high point, the low point is connected with the low point, and then K line shape classification is carried out according to the relation of the high and low points of the K line phase to form a K line shape library;

the K-line pattern morphology can be divided into four morphologies, greater than, less than, and upward and downward.

b1. Class > as shown in fig. 2: a > -C, B < ═ D;

for example, 000568 Luzhou Lao jiao,

a: time: 2020/01/13 closing price: 93.03

B: time: 2020/02/04 closing price: 72.12

C: time: 2020/02/20 closing price: 82.77

D: time: 2020/02/28 closing price: 74.80

b2. Less than class, as shown in FIG. 3: a < C, B > D;

for example, 600519 Guizhou Maotai,

a: time: 2020/02/20 closing price: 1118

B: time: 2020/02/28 closing price: 1057

C: time: 2020/03/05 closing price: 1171

D: time: 2020/03/19 closing price: 996

b3. Upward class, as shown in FIG. 4: a < C, B < ═ D;

for example, 600585A cement for conchs,

a: time: 2020/01/02 closing price: 55.80

B: time: 2020/02/03 closing price: 44.83

C: time: 2020/03/02 closing price: 60.31

D: time: 2020/03/19 closing price: 50.09

b4. Downward, as shown in FIG. 5: a > -C, B > D.

For example, 601318 is safe in China,

a: time: 2020/02/20 closing price: 83.28

B: time: 2020/02/28 closing price: 77.72

C: time: 2020/03/05 closing price: 82.45

D: time: 2020/03/23 closing price: 66.76

(4) The system sets the K line state to be found, traverses the K line state library, and finds out the corresponding K line state from a plurality of K lines in the K line state library.

Claims

1. A K-ray shape recognition method is characterized by comprising the following steps:

(1) accessing a financial market data source;

(2) finding out high and low points of the closing price in the K line stage by using a ZIG function, and respectively defining the high and low points as a next-new high point, a next-new low point, a new high point and a new low point;

(3) connecting high and low points of adjacent stages, and classifying K-line states according to the relation of the high and low points of the K-line stage to form a K-line state library; the method for connecting the high points and the low points of the adjacent stages comprises the following steps: connecting a high point and a high point, and connecting a low point and a low point; when the K line form is classified, the sizes of the front and rear high and low points are judged to realize form classification;

(4) setting a K line state to be found, traversing a K line state library, finding a corresponding K line state in the K line state library, dividing the K line state into four first-order forms of a higher form, a lower form, an upper form and a lower form, and if a next new high point, a next new low point, a new high point and a new low point are respectively recorded as A, B, C, D, the higher form is expressed as: a > ═ C, B < ═ D; the less than class morphology is represented as: a < C, B > D; the upward morphology is represented as: a < C, B < ═ D; the downward morphology is represented as: a > -C, B > D.