Advanced Clustering Technique For High-Frequency Low-Latency Analysis
After configuring the backtesting scenario, data is streamed to ExSan to identify clusters that will subsequently be utilized for constructing the covariance matrix. This process is fundamental to modern portfolio theory, particularly due to the asynchronous nature of market data. Only clusters with overlapping time periods (i.e., concurrent in time) are eligible for inclusion in the estimation of the covariance matrix.
The aim of this work is to establish Covariance and Correlation of Asynchronous Market Tick Data Using Cluster Computing. An algorithm is implemented to correlate asynchronous tick data from the market using cluster computing. Each asset generates a unique number of ticks distributed over a given time frame, with some assets being more active and others less so, reflecting varying levels of volatility and activity.
Challenges Addressed
Performing some kind of extreme test under strict timing.
Total Number of assets allocated: 40 clusters: 20 per asset(stock) time lapse: 3333 microseconds
OLD Version Cluster NOT DECLARED STATIC
xsnPortfolio_0222 Elapsed Time: 161 sec
xsnPortfolio_0627 Elapsed Time: 30 sec
xsnPortfolio_0906 Elapsed Time: 748 sec
xsnPortfolio_3115 Elapsed Time: 484 sec
xsnPortfolio_4504 Elapsed Time: 403 sec
New Version Cluster DECLARED STATIC !
xsnPortfolio_2903 Elapsed Time: 1.40 sec
xsnPortfolio_2927 Elapsed Time: 2.93 sec
xsnPortfolio_2951 Elapsed Time: 4.42 sec
xsnPortfolio_3020 Elapsed Time: 2.75 sec
xsnPortfolio_3045 Elapsed Time: 2.79 sec
xsnPortfolio_4022 Elapsed Time: 4.24 sec
xsnPortfolio_4112 Elapsed Time: 12.01 sec
xsnPortfolio_4304 Elapsed Time: 6.21 sec
Pseudo Code
tau <- Set time lag
startChrono
while endChrono - startChrono < tau
feed Data Market to ExSan into current cluster
endChrono
ExSan Abstract Structure
Clusters
ExSan + Clusters
Remark: All clusters are implemented as Red-Black Trees to ensure consistency with the earlier description of ExSan. This implementation leverages a single C++ master template, designed and implemented only once, eliminating the need for additional coding. The master template manages all cluster-related functionality, ensuring uniformity, efficiency, and maintainability across the system, which is central to ExSan's architecture.
ExSan was conceived 35 years ago, when it first started as a coding project. The original version was written in C and was later rewritten in C++. From the very beginning, the project was based on Red-Black Trees.
This LinkedIn post highlights the significance of Red-Black Trees as a foundational component in a series of articles focused on Quantitative Finance. During ExSan’s conception and development, it was not evident that similar concepts utilizing Red-Black Trees were being explored by others. This underscores the independent and innovative thought process behind ExSan’s creation.
Output
1 |ExSan| C++ |ExSan| Sun Dec 15 08:51:29 2024
2 JOB: xsnPortfolio_5129
3 <--------e-x-s-a-n---------->
4 Generate Exsan ( 23 , 24 )
5 Total Number of assets allocated: 22 Lapse Microseconds: 7
6 |<--------e-x-s-a-n---hft------->|
7 :2291: Catching Clusters This round ExSan(stocks: 22, 2 :clusters)
8 :2292: Selected Stocks |aryStockAux| Weighted Vals in Clusters ppDataIn: 9
9 WORKSHEET 9 I[5, 2] FLOAT
10 A B
11 >--------------<
12 0 0.00 ATLU 2:272.95 278.50
13 0 0.00 ARNA 4: 62.00 63.20
14 0 0.00 ANRA 5: 11.50 11.50
15 <-------------->
16 1 2
17 :2321: CATCHING CLUSTERS Clocking - Time Lapse Range
18 ATLU 1 [10.99 <-> 10.99]2 [11.30 <-> 12.03 ]
19 ARNA 1 [11.07 <-> 11.07]2 [11.42 <-> 11.42]
20 ANRA 1 [11.12 <-> 11.12 ]2 [12.00 <-> 12.00]
21
22 *** The time range of Cluster j, defined as [min, max], falls within the overall time window ***
23 [10.99 <-> 11.12 ] [11.30 <-> 12.03 ]
24
25 ENDS xsnPortfolio_5129 Elapsed Time: 0.04 sec
26 Version BOOST: 1.83.0 EXSAN @ MS VSC 2022 (64b) - 17.11.5 -toolSet(v143)
27 EXIT FROM EXSAN
28
29 *************************************************************************
30 |ExSan| C++ |ExSan| Sun Dec 15 08:55:56 2024
31 JOB: xsnPortfolio_5556
32 <--------e-x-s-a-n---------->
33 Generate Exsan ( 23 , 24 )
34 Total Number of assets allocated: 22 Lapse Microseconds: 10
35 |<--------e-x-s-a-n---hft------->|
36 :2291: Catching Clusters This round ExSan(stocks: 22, 3 :clusters)
37 :2292: Selected Stocks |aryStockAux| Weighted Vals in Clusters ppDataIn: 9
38 WORKSHEET 9 I[14, 3] FLOAT
39 A B C
40 >---------------------<
41 0 0.00 ATHN 5: 27.76 31.13 31.37
42 0 0.00 AKSH 6: 97.88 103.79 106.30
43 0 0.00 ARNA 7: 74.60 73.40 71.70
44 0 0.00 ANRA 8: 13.40 11.50 11.90
45 0 0.00 ADVN 10:139.33 144.56 151.75
46 0 0.00 ARAC 12: 12.93 12.47 12.84
47 0 0.00 ALV 13: 58.05 55.07 55.22
48 0 0.00 ANUL 14: 10.90 10.88 10.78
49 <--------------------->
50 1 2 3
51
52 :2321: CATCHING CLUSTERS Clocking - Time Lapse Range
53 ATHN 1 [
10.43
<-> 10.55]2 [
11.31
<->
12.23
]3 [12.81 <-> 12.81]
54 AKSH 1 [10.51 <-> 10.51]2 [11.56 <-> 12.04]3 [13.13 <-> 13.13]
55 ARNA 1 [10.59 <-> 11.16]2 [11.89 <-> 11.89]3 [12.50 <-> 12.92]
56 ANRA 1 [10.72 <->
11.23
]2 [11.68 <-> 11.68]3 [12.74 <-> 12.96]
57 ADVN 1 [10.84 <-> 10.84]2 [11.75 <-> 11.75]3 [
12.43
<-> 12.84]
58 ARAC 1 [10.96 <-> 10.96]2 [12.08 <-> 12.08]3 [12.47 <-> 13.10]
59 ALV 1 [11.02 <-> 11.19]2 [12.01 <-> 12.01]3 [12.67 <->
13.16
]
60 ANUL 1 [11.07 <-> 11.13]2 [11.79 <-> 11.98]3 [12.56 <-> 13.03]
61
62 *** The time range of Cluster j, defined as [min, max], falls within the overall time window ***
63
[10.43 <-> 11.23]
[11.31 <-> 12.23]
[12.43 <-> 13.16]
64
65 ENDS xsnPortfolio_5556 Elapsed Time: 0.04 sec
66 Version BOOST: 1.83.0 EXSAN @ MS VSC 2022 (64b) - 17.11.5 -toolSet(v143)
67 EXIT FROM EXSAN
1 |ExSan| C++ |ExSan| Sun Dec 15 08:51:29 2024
2 JOB: xsnPortfolio_5129
3 <--------e-x-s-a-n---------->
4 Generate Exsan ( 23 , 24 )
5 Total Number of assets allocated: 22 Lapse Microseconds: 7
6 |<--------e-x-s-a-n---hft------->|
7 :2291: Catching Clusters This round ExSan(stocks: 22, 2 :clusters)
8 :2292: Selected Stocks |aryStockAux| Weighted Vals in Clusters ppDataIn: 9
9 WORKSHEET 9 I[5, 2] FLOAT
10 A B
11 >--------------<
12 0 0.00 ATLU 2:272.95 278.50
13 0 0.00 ARNA 4: 62.00 63.20
14 0 0.00 ANRA 5: 11.50 11.50
15 <-------------->
16 1 2
17 :2321: CATCHING CLUSTERS Clocking - Time Lapse Range
18 ATLU 1 [10.99 <-> 10.99]2 [11.30 <-> 12.03]
19 ARNA 1 [11.07 <-> 11.07]2 [11.42 <-> 11.42]
20 ANRA 1 [11.12 <-> 11.12]2 [12.00 <-> 12.00]
21
22 *** The time range of Cluster j, defined as [min, max], falls within the overall time window ***
23 [10.99 <-> 11.12] [11.30 <-> 12.03]
24
25 ENDS xsnPortfolio_5129 Elapsed Time: 0.04 sec
26 Version BOOST: 1.83.0 EXSAN @ MS VSC 2022 (64b) - 17.11.5 -toolSet(v143)
27 EXIT FROM EXSAN
28
29
30 *************************************************************************
31 |ExSan| C++ |ExSan| Sun Dec 15 08:55:56 2024
32 JOB: xsnPortfolio_5556
33 <--------e-x-s-a-n---------->
34 Generate Exsan ( 23 , 24 )
35 Total Number of assets allocated: 22 Lapse Microseconds: 10
36 |<--------e-x-s-a-n---hft------->|
37 :2291: Catching Clusters This round ExSan(stocks: 22, 3 :clusters)
38 :2321: CATCHING CLUSTERS Clocking - Time Lapse Range
39 ATHN 1 [10.43 <-> 10.55]2 [11.31 <-> 12.23]3 [12.81 <-> 12.81]
40 AKSH 1 [10.51 <-> 10.51]2 [11.56 <-> 12.04]3 [13.13 <-> 13.13]
41 ARNA 1 [10.59 <-> 11.16]2 [11.89 <-> 11.89]3 [12.50 <-> 12.92]
42 ANRA 1 [10.72 <-> 11.23]2 [11.68 <-> 11.68]3 [12.74 <-> 12.96]
43 ADVN 1 [10.84 <-> 10.84]2 [11.75 <-> 11.75]3 [12.43 <-> 12.84]
44 ARAC 1 [10.96 <-> 10.96]2 [12.08 <-> 12.08]3 [12.47 <-> 13.10]
45 ALV 1 [11.02 <-> 11.19]2 [12.01 <-> 12.01]3 [12.67 <-> 13.16]
46 ANUL 1 [11.07 <-> 11.13]2 [11.79 <-> 11.98]3 [12.56 <-> 13.03]
47
48 *** The time range of Cluster j, defined as [min, max], falls within the overall time window ***
49 [10.43 <-> 11.23] [11.31 <-> 12.23] [12.43 <-> 13.16]
50
51 ENDS xsnPortfolio_5556 Elapsed Time: 0.04 sec
52 EXIT FROM EXSAN
53
54
55 *************************************************************************
56 |ExSan| C++ |ExSan| Sun Dec 15 08:55:56 2024
57 JOB: xsnPortfolio_5556
58 :2321: CATCHING CLUSTERS Clocking - Time Lapse Range
59 ANUL 1[ 9.74 <-> 10.04] 2[10.43 <-> 11.01] 3[11.66 <-> 11.96] 4[12.96 <-> 13.19] 5[13.88 <-> 13.88]
60 ASCMA 1[ 9.89 <-> 9.89] 2[11.25 <-> 11.25] 3[11.80 <-> 11.80] 4[13.16 <-> 13.16] 5[13.39 <-> 13.72]
61 ADVN 1[ 9.96 <-> 9.96] 2[10.46 <-> 10.99] 3[11.47 <-> 12.23] 4[12.63 <-> 12.63] 5[13.50 <-> 13.93]
62 AMRN 1[10.13 <-> 10.19] 2[11.05 <-> 11.05] 3[12.13 <-> 12.13] 4[12.86 <-> 12.90] 5[14.19 <-> 14.19]
63
64 *** The time range of Cluster j, defined as [min, max], falls within the overall time window ***
65 [9.74 <-> 10.19] [10.43 <-> 11.25] [11.47 <-> 12.23] [12.63 <-> 13.19] [13.39 <-> 14.19]
66 ENDS xsnPortfolio_4449 Elapsed Time: 0.04 sec
67
68
69 ----MORE RESULTS---
70 *************************************************************************
71 :2321: CLUSTERS CATCHING Clocking - Time Lapse Range
74
75 *** The time range of Cluster j, defined as [min, max], falls within the overall time window ***
77
78 EXIT FROM EXSAN
79 *************************************************************************
80 :2321: CLUSTERS CATCHING Clocking - Time Lapse Range
81 ASCMA 4[215.40 <-> 215.40] 5[216.37 <-> 216.37] 1[217.81 <-> 217.81] 2[218.35 <-> 218.95] 3[219.88 <-> 219.88]
82 CAR 2[215.83 <-> 216.02] 3[216.70 <-> 216.70] 4[217.33 <-> 217.41] 5[218.50 <-> 218.62] 1[220.01 <-> 220.01]
83 ADVN 3[215.24 <-> 215.24] 4[216.95 <-> 216.95] 5[217.38 <-> 217.38] 1[218.87 <-> 218.87] 2[219.96 <-> 219.96]
84 BMY 1[215.28 <-> 215.28] 2[216.42 <-> 216.58] 3[217.78 <-> 217.78] 4[218.43 <-> 218.69] 5[219.39 <-> 219.39]
85
86 *** The time range of Cluster j, defined as [min, max], falls within the overall time window ***
87 [215.24 <-> 216.02] [216.37 <-> 216.95] [217.33 <-> 217.81] [218.35 <-> 218.95] [219.39 <-> 220.01]
88 EXIT FROM EXSAN
89 *************************************************************************
90 :2321: CLUSTERS CATCHING Clocking - Time Lapse Range
91 CAR 5[21671.61 <-> 21671.61] 1[21672.59 <-> 21672.59] 2[21673.81 <-> 21673.81] 3[21674.74 <-> 21674.74] 4[21676.42...]
92 BF 2[21671.70 <-> 21671.70] 3[21673.39 <-> 21673.39] 4[21674.21 <-> 21674.21] 5[21675.02 <-> 21675.02] 1[21675.65...]
93 AKS 4[21671.77 <-> 21671.97] 5[21672.56 <-> 21672.89] 1[21673.65 <-> 21673.65] 2[21675.35<->21675.35] 3[21676.05<->...]
94 ARNA 1[21672.09 <-> 21672.09] 2[21672.70 <-> 21673.10] 3[21674.29 <-> 21674.29] 4[21674.94 <-> 21674.94] 5[21675.86<...]
95 AAL 5[21672.18 <-> 21672.18] 1[21672.77 <-> 21672.77] 2[21674.25 <-> 21674.25] 3[21674.55 <-> 21675.32] 4[21676.29<...]
96 BLGS 5[21672.28 <-> 21672.32] 1[21673.18 <-> 21673.18] 2[21673.78 <-> 21673.78] 3[21674.63 <-> 21674.63] 4[21676.13<...]
97 C 3[21671.66 <-> 21671.66] 4[21673.01 <-> 21673.01] 5[21674.33 <-> 21674.33] 1[21674.90 <-> 21675.27] 2[21675.82<...]
98
99 *** The time range of Cluster j, defined as [min, max], falls within the overall time window ***
100 [21671.61 <-> 21672.32] [21672.56 <-> 21673.39] [21673.65 <-> 21674.33] [21674.55 <-> 21675.35] [21675.65...
101 NO OVERLAPPING TIME FRAMES, ALL CLUSTER FIT PROPERLY
101 EXIT FROM EXSAN
102 *************************** E N D *****************************************
103
Traversing data for each cluster
1
2 |ExSan| C++ |ExSan| Mon Dec 16 07:27:53 2024
3
4 JOB: xsnPortfolio_2753
5 <--------e-x-s-a-n---------->
6 Generate Exsan ( 7 , 8 )
7 CLUSTERING ENABLE
8 microSecTimeLapse: 1000 Lapse Microseconds: 1000 minTics: 77777 maxTics: 77777
9 |<--------e-x-s-a-n---hft------->|
10
11 I[6, 3] *Matrix of Weighted Clusters*
12 A B C
13 >---------------------<
14 0 0.00 TFSM 2: 71.77 69.80 71.21
15 0 0.00 MSFT 3: 24.40 23.09 23.47
16 0 0.00 AAPL 4: 22.30 21.45 21.18
17 0 0.00 XLFN 6:158.75 146.25 158.02
18 <--------------------->
19 1 2 3
20
21 :2321: CLUSTERS CATCHING Clocking - Time Lapse Range
22 TFSM 1[9.95 <-> 10.65] 2[10.98 <-> 11.74] 3[11.90 <-> 13.56]
23 (2 ,1 ):
24 pp: 9 71.77 ->
25 1 2024-Dec-16 07:27:53.294270 pnt: 0 71.77
26 2 2024-Dec-16 07:27:53.294270 pnt: 0 71.15
27 4 2024-Dec-16 07:27:53.294270 pnt: 0 69.99
28 6 2024-Dec-16 07:27:53.294270 pnt: 0 69.98
29 8 2024-Dec-16 07:27:53.294270 pnt: 0 69.46
30
31 (2 ,2 ):
32 pp: 9 69.80 ->
33 1 2024-Dec-16 07:27:53.295271 pnt: 0 68.57
34 3 2024-Dec-16 07:27:53.295271 pnt: 0 69.08
35 5 2024-Dec-16 07:27:53.295271 pnt: 0 68.17
36 7 2024-Dec-16 07:27:53.295271 pnt: 0 68.93
37 9 2024-Dec-16 07:27:53.295271 pnt: 0 68.49
38 11 2024-Dec-16 07:27:53.295271 pnt: 0 69.80
39
40 (2 ,3 ):
41 pp: 9 71.21 ->
42 1 2024-Dec-16 07:27:53.296276 pnt: 0 69.21
43 3 2024-Dec-16 07:27:53.296276 pnt: 0 70.53
44 5 2024-Dec-16 07:27:53.296276 pnt: 0 71.21
45 7 2024-Dec-16 07:27:53.296276 pnt: 0 70.26
46 9 2024-Dec-16 07:27:53.296276 pnt: 0 70.27
47 11 2024-Dec-16 07:27:53.297269 pnt: 0 69.91
48 13 2024-Dec-16 07:27:53.297269 pnt: 0 70.87
49 15 2024-Dec-16 07:27:53.297269 pnt: 0 70.00
50 17 2024-Dec-16 07:27:53.297269 pnt: 0 69.90
51
52 MSFT 1[10.00 <-> 10.28] 2[11.38 <-> 11.78] 3[11.99 <-> 13.74]
53 (3 ,1 ):
54 pp: 9 24.40 ->
55 1 2024-Dec-16 07:27:53.294270 pnt: 0 24.40
56 2 2024-Dec-16 07:27:53.294270 pnt: 0 24.10
57 4 2024-Dec-16 07:27:53.294270 pnt: 0 24.02
58 6 2024-Dec-16 07:27:53.294270 pnt: 0 23.45
59 8 2024-Dec-16 07:27:53.294270 pnt: 0 22.99
60
61 (3 ,2 ):
62 pp: 9 23.09 ->
63 1 2024-Dec-16 07:27:53.295271 pnt: 0 23.00
64 3 2024-Dec-16 07:27:53.295271 pnt: 0 22.68
65 5 2024-Dec-16 07:27:53.295271 pnt: 0 22.92
66 7 2024-Dec-16 07:27:53.295271 pnt: 0 23.09
67 9 2024-Dec-16 07:27:53.295271 pnt: 0 22.66
68
69 (3 ,3 ):
70 pp: 9 23.47 ->
71 1 2024-Dec-16 07:27:53.296276 pnt: 0 23.11
72 3 2024-Dec-16 07:27:53.296276 pnt: 0 23.12
73 5 2024-Dec-16 07:27:53.297269 pnt: 0 23.47
74 7 2024-Dec-16 07:27:53.297269 pnt: 0 22.92
75 9 2024-Dec-16 07:27:53.297269 pnt: 0 22.46
76 11 2024-Dec-16 07:27:53.297269 pnt: 0 22.40
77 13 2024-Dec-16 07:27:53.297269 pnt: 0 22.19
78 15 2024-Dec-16 07:27:53.297269 pnt: 0 22.06
79
80 AAPL 1[10.04 <-> 10.80] 2[11.17 <-> 11.26] 3[11.87 <-> 13.60]
81 (4 ,1 ):
82 pp: 9 22.30 ->
83 1 2024-Dec-16 07:27:53.294270 pnt: 0 22.30
84 2 2024-Dec-16 07:27:53.294270 pnt: 0 22.17
85 4 2024-Dec-16 07:27:53.294270 pnt: 0 21.31
86 6 2024-Dec-16 07:27:53.294270 pnt: 0 21.46
87 8 2024-Dec-16 07:27:53.295271 pnt: 0 20.90
88
89 (4 ,2 ):
90 pp: 9 21.45 ->
91 1 2024-Dec-16 07:27:53.295271 pnt: 0 21.45
92 3 2024-Dec-16 07:27:53.295271 pnt: 0 20.52
93
94 (4 ,3 ):
95 pp: 9 21.18 ->
96 1 2024-Dec-16 07:27:53.296276 pnt: 0 20.92
97 3 2024-Dec-16 07:27:53.296276 pnt: 0 19.92
98 5 2024-Dec-16 07:27:53.296276 pnt: 0 20.27
99 7 2024-Dec-16 07:27:53.296276 pnt: 0 20.63
100 9 2024-Dec-16 07:27:53.297269 pnt: 0 20.51
101 11 2024-Dec-16 07:27:53.297269 pnt: 0 20.21
102 13 2024-Dec-16 07:27:53.297269 pnt: 0 19.77
103 15 2024-Dec-16 07:27:53.297269 pnt: 0 20.73
104 17 2024-Dec-16 07:27:53.297269 pnt: 0 21.18
105 19 2024-Dec-16 07:27:53.297269 pnt: 0 21.10
106
107 XLFN 1[10.33 <-> 10.71] 2[11.36 <-> 11.36] 3[12.05 <-> 13.78]
108 (6 ,1 ):
109 pp: 9 158.75 ->
110 1 2024-Dec-16 07:27:53.294270 pnt: 0 158.75
111 2 2024-Dec-16 07:27:53.294270 pnt: 0 154.33
112 4 2024-Dec-16 07:27:53.294270 pnt: 0 152.67
113 6 2024-Dec-16 07:27:53.294270 pnt: 0 150.18
114
115 (6 ,2 ):
116 pp: 9 146.25 ->
117 1 2024-Dec-16 07:27:53.295271 pnt: 0 146.25
118
119 (6 ,3 ):
120 pp: 9 158.02 ->
121 1 2024-Dec-16 07:27:53.296276 pnt: 0 147.61
122 3 2024-Dec-16 07:27:53.296276 pnt: 0 146.17
123 5 2024-Dec-16 07:27:53.296276 pnt: 0 149.41
124 7 2024-Dec-16 07:27:53.296276 pnt: 0 150.09
125 9 2024-Dec-16 07:27:53.296276 pnt: 0 153.41
126 11 2024-Dec-16 07:27:53.297269 pnt: 0 151.03
127 13 2024-Dec-16 07:27:53.297269 pnt: 0 157.50
128 15 2024-Dec-16 07:27:53.297269 pnt: 0 158.02
129 17 2024-Dec-16 07:27:53.297269 pnt: 0 154.89
130 19 2024-Dec-16 07:27:53.297269 pnt: 0 155.03
131 21 2024-Dec-16 07:27:53.298268 pnt: 0 152.05
132
133
134 *** The time range of Cluster j, defined as [min, max], falls within the overall time window ***
135 [9.95 <-> 10.80] [10.98 <-> 11.78] [11.87 <-> 13.78]
136
137 ENDS xsnPortfolio_2753 Elapsed Time: 0.05 sec
138 EXIT FROM EXSAN
Some Pseudo Code of ExSan Cluster
IF loadNtimeLapse == 1 THEN
// Use auxiliary tree to process data
Initialize ndptrAry[0] and ndptrAry[1] with ptrAry[0]'s ntHead
Find minimum value in the tree and assign to ndptrAry[0]
Find maximum value in the tree and assign to ndptrAry[1]
Calculate spreadCenter as the average of minimum and maximum data
IF ntCounter is odd THEN
Calculate kth as middle index of the tree
Find kth node using InOrderFindKth_nt
Set data and time from the kth node's data
ELSE
Calculate kth for left and right midpoints
Find nodes for left and right midpoints using InOrderFindKth_nt
Set data and time as the average of left and right midpoint nodes
ENDIF
Discard the auxiliary tree to free resources
Validate tree cleanup and output results if required
ELSE
// Handle new version without deleting nodes
Initialize ndptrAry[0] and ndptrAry[1] to null
Find minimum and maximum keys using rangeKeyFindMinMax_nt
Calculate spreadCenter as the average of minimum and maximum data
IF key_index is odd THEN
Find the middle node using rangeKeyFind_nt
Set data and time from the middle node's data
ELSE
Find left and right midpoint nodes using rangeKeyFindPair_nt
Set data and time as the average of the two midpoint nodes
ENDIF
ENDIF
Some ExSan Cluster Code
1
2
3 /*Some ExSan Code Clocking Clusters*/
4 fout << "\n\t:2321: CATCHING CLUSTERS Clocking - Time Lapse Range ";
5 stockPtr = net->goto_row(net, ppDataIn, firstRow, 'f');
6 for (i = firstRow; i <= lastRow; ++i) {
7 if (!aryStockAux[i]) { stockPtr = stockPtr->get_down_ptr(); continue; }
8 ptr = stockPtr->get_forward_ptr();
9
10 cout << "\n\t" << setw(5) << left << stockPtr->get_stock_name() << "\t";
11
12 for (j = 1; j <= net->get_cols_in_page(ppDataIn, 'f'); j++) { /
13 if (loadNtimeLapse == 1) {
14 ndptrAry[0] = ndptrAry[1] = nullptr;
15 ptrAry[0]->get_ntHead()->rangeKeyFindMinMax_nt( ... /* some parameters*/... );
16 cout << setw(2) << ptrAry[0]->get_col() << "[" << ...
17
18 if (ndptrAry[0]->get_nt_data(ppnt_time) < ptrAry[0]->get_data(ppTime))
19 ptrAry[0]->set_data(ppTime, ndptrAry[0]->get_nt_data(ppnt_time));
20 if (ndptrAry[1]->get_nt_data(ppnt_time) > ptrAry[0]->get_data(ppRTime))
21 ptrAry[0]->set_data(ppRTime, ndptrAry[1]->get_nt_data(ppnt_time));
22 }
23 else {/
24 ndptrAry[0] = ndptrAry[1] = nullptr;
27 << " <-> " << setw(7) << ndptrAry[1]->get_nt_data(ppnt_time) << "]"
28 }
29 ptrAry[0] = ptrAry[0]->get_next_ptr();
30 if (ptrAry[0]->get_col() > net->get_cols_in_page(ppDataIn, 'f')) ptrAry[0] = ....;
31 }
32 }
33
34 ptr = net->net_point_to(net, net->get_net_rows(), 1);
35
36 for (j = 1; j <= net->get_cols_in_page(ppDataIn, 'f'); j++) { /
37 ptr->set_data(ppTime, ptrAry[0]->get_data(ppTime));
38 ptr->set_data(ppRTime, ptrAry[0]->get_data(ppRTime));
39 ptrAry[0] = ptrAry[0]->get_next_ptr();
40 if (ptrAry[0]->get_col() > net->get_cols_in_page(ppDataIn, 'f'))
41 ptrAry[0] = stockPtr->get_forward_ptr();
42 ptr = ptr->get_next_ptr();
43 }
44 cout << endl;
45
46 stockPtr = net->goto_row(net, ppTime, firstRow, 'f'); /
48 if (!aryStockAux[i]) {
49 stockPtr = stockPtr->get_down_ptr();
50 continue;
51 }
52
53 ptrAry[0] = stockPtr->get_ptrLasTick(); /
54 ptr = net->net_point_to(net, net->get_net_rows(), 1);
55 for (j = 1; j <= net->get_cols_in_page(ppTime, 'f'); j++) { /
56
57 if(ptrAry[0]->get_data(ppTime) < ptr->get_data(ppTime)) {
58 ptr->set_data(ppTime, ptrAry[0]->get_data(ppTime));
59 }
60 if(ptrAry[0]->get_data(ppRTime) > ptr->get_data(ppRTime)) {
61 ptr->set_data(ppRTime, ptrAry[0]->get_data(ppRTime));
62 }
63
64 ptrAry[0] = ptrAry[0]->get_next_ptr();
65 if (ptrAry[0]->get_col() > net->get_cols_in_page(ppDataIn, 'f'))
66 ptrAry[0] = stockPtr->get_forward_ptr();
67 ptr = ptr->get_next_ptr();
68 }
69 stockPtr = stockPtr->get_down_ptr();
70 }
71
73 for (j = 1; j <= net->get_cols_in_page(ppTime, 'f'); ++j) { /
75 ptr = ptr->get_next_ptr();
76 }
77 cout endl;
Some Outputs
ExSan's Quant Finance Backtesting Suite: Performance & Outcomes
Comments
Post a Comment