Predicting investment fluxes from implicit lead-lag investor … · 2019-01-10 · Predicting...
Transcript of Predicting investment fluxes from implicit lead-lag investor … · 2019-01-10 · Predicting...
Predicting investment fluxesfrom implicit lead-lag investor networks
D. Challet1,2 M. Lallouache 1 R. Chicheportiche 3,4
1 CentraleSupelec 2 Encelade Capital SA 3CFM 4 previously at Swissquote Bank SA
September 8, 2015
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Broker internal order matching
versus
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Internal order matching
...
?
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Broker’s gain
GT =T
∑t=1
Itrt =T
∑t=1
(It−1+δ It)rt
Classic optimization problem1 Fix T (1 day)2 Assume random δ It and rt
3 Constraints4 Minimize cost function
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Broker’s gain
GT =T
∑t=1
Itrt =T
∑t=1
(It−1+δ It)rt
Prediction problem1 predict flux2 predict price return3 PROFIT!
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Prediction problem= Science + Engineering
Science: flux1 cluster clients2 lead-lag networks3 machine learning4 ENJOY!
Engineering1 Predict price returns2 Inventory constraints3 PROFIT!
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Investor classification
SupervisedIndividual vs institutional (Odean, Barber, etc.)
...
UnsupervisedSimilarity measure
→ categories
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Supervised classification
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Real networks
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Unsupervised learning
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Unsupervised learning
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
FX traders
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●●
●
●●
●
●●
●
●
●
●●●●●
●
●
●
●
●
●
● ●
●●
●●
●
●●
●
●
●
●
●
●●
●
●
●
●
●
●●
●
●●●
●
●
●
●
●
●
●
●●
●●
●●
●●
●
●
●
●●
●
●
●
●
●●
●●●●●
●●●
●●●●
●●
●
●
●
●●
●
●
●
●
●
●
●
●
50 100 150 200
−8
−6
−4
−2
02
yday
cum
sum
(str
at)
●●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●●
●
●
●●
●
●
●●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●●
●●
●
●
●●●
●
●
●
●
●
●
●●●
●
●
●●●
●
●
●
●
●
●
●
●●●
●●●
50 100 150 200
−10
−5
05
10
yday
cum
sum
(str
at)
●●●
●
●
●●
●
●
●
●
●
●●
●●
●
●
●
●●
●
●
●
●●
●●●
●
●
●
●●●
●●●
●
●
●
●●●
●
●
●●●
●
●●
●●
●●●●
●
●●●●●●●●●
●●●
●
●●
●●
●
●
●●
●●●●●●●
●
●
●
●●●
●●
●
●
●
●
●
●●●●
●
●●●
●
●●●●●●
●
●●●●●
●●●
●
●●
●
●●●●●●●●●
50 100 150 200
−2
02
46
8
yday
cum
sum
(str
at)
●●●
●●●●●
●●●●●●
●●●●
●●●●●
●●●●●●
●●●●
●●●●●
●●●●●
●●●●●
●●●●●
●●●●● ●
●●●
●●●●●
●●●●●
●●●●●
●●●●●●
●●●●
●●●●●
●●●●●
●●●●●
●●●●●
●●●●●
●●●●●
●●●●●
●●●●●
●●●●●
●●●●●
50 100 150 200
010
2030
4050
yday
cum
sum
(str
at)
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Unsupervised classification ISpanish brokers
Lillo et al. (2008)
Daily inventory change Vi(t)
Correlation matrixC ∼ E(ViVj)
Principal ComponentAnalysis + Random Matrixtheory
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Unsupervised classification IEquities, Spanish brokers
Lillo et al. (2008)
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Unsupervised classification IIEquities, Chinese investors
Zhou et al. (2012)
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Unsupervised classification III:FX, individual investors, daily
λ
p(λ)
0 2 4 6 8
0.0
0.2
0.4
0.6
0.8
1.0 Data
ShuffledRMT
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●●
500 1000 2000 5000 10000−
0.8
−0.
6−
0.4
−0.
20.
00.
20.
4
log(N)
CS
R
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Unsupervised classification IV:FX, individual investors, hourly
λ
p(λ)
0 2 4 6 8 10
0.0
0.2
0.4
0.6
0.8
1.0 Data
ShuffledRMT
●
●
●
●
●●
●
●●
●
●
●
●
●
●●
●
●
● ●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●● ●
●
●
●
●●
●
●
●●
●
●
●
500 1000 2000 5000 10000 20000 50000−
0.6
−0.
4−
0.2
0.0
0.2
0.4
log(N)
CS
R
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Unsupervised classificationStatistically validated networks (SVNs)
Tumminello et al. (2011a)
1 Agent i, statei(t) ∈ {1, · · · ,S}
si(t) = signBuy(t)−Sell(t)Buy(t)+Sell(t)
∈ {−1,0,1,2 = /0}
Agents i and j: measure frequency
si(t) = s && sj(t) = s′
Compute p-values
O(N2) pairs→ multiple hypothesis correction→NETWORK
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Origin of trade synchronicity
1 Explicit communication2 Implicit communication
1 same news2 same strategy: MA(CD)3 same parameters4 Master in Finance
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Trader-trader communication: explicit
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Trader-trader communication: explicit
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Trader-trader communication: explicit
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Do traders look at prices?
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Implicit AND explicit communication
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Implicit communication: same price analysis
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Implicit communication: same price analysis
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Implicit communication: same price analysis
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Example: MACD (1970)
Parameters 12, 26, and 9 days
Trading week then: 6 days
Now: 5 days
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Investor SVNsgroup=disconnected subgraphs
Tumminello et al. (2011b): daily Finish investment fluxes
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Datasets
1 Swissquote (SQ): individual traders2 Large Bank: institutional traders
10-3
10-2
10-1
100
100 101 102 103 104
N
P >( N)
10-4
10-3
10-2
10-1
100
100 101 102 103 104 105
N
P >( N)
Number of transactions
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Step 1: Implicit network + community, daily
Challet et al. (2015) EUR.USD SQ
●
●
●
●
●
●
●
●
●●●
●
●●
●
●●
●●●
●●
●
●●
●●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●●
● ●
●
●
●●
●●
●
●●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●●
Links
−1 −11 11 22 2
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Implicit network + community, hourly
Challet et al. (2015) EUR.USD SQ
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●● ●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●●
●●
●
●●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
● ●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
Links
−1 −11 −12 −1−1 11 12 1−1 21 22 2
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Implicit network + community detection, LB, hourly
Challet et al. (2015)
GBP.USD EUR.USD
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●Links
1 1−1 −1−1 11 −1
● ●
●
●
● ●
●
●
●
●
●
●
●●
●
●●
●
●
●●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●● ●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
Links
1 −11 1−1 −1−1 12 1−1 22 −11 22 2
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Cluster stability
1_12_1
1_2
1_3
1_15
1_24
1_16
1_17
1_26
1_29
2_37
1_4 2_2
2_3
1_31
1_7
2_8
1_10
2_9
1_8
2_5
2_21
2_10
1_21
1_23
2_27
1_11
2_12
1_222_26
2_19
2_36
2_22
2_25
1_20
1_27
2_32
1_30
2_34
2_24
2_28
1_28
2_33
2_29
2_31
1_0
2_17
2_182_20
2_23
2_0
3_1
3_2
3_3
3_5
3_7
3_13
3_29
3_10
3_20
3_30
3_14
3_34
3_21
3_27
3_39
3_15
3_19
3_40
3_31
3_37
3_16
3_42
3_32
3_38
3_26
3_28
3_35
3_12
3_25
3_243_22
3_36
4_1
4_3
4_2
4_6
4_9
4_13
4_22
4_15
4_25
4_7
4_34
4_14
4_23
4_24
4_30
4_19
4_28
4_33
4_37
4_27
4_11
4_32
4_35
4_36
4_314_17
4_21
4_26
4_0
5_1
5_2
5_3
5_7
5_16
5_11
5_30
5_8
5_34
5_10
5_36
5_32
5_29
5_14
5_42
5_47
5_21
5_40
5_31
5_26
5_275_22
5_24
5_44
5_48
5_35
5_28
5_37
5_15
5_18
5_45
5_12
5_19
5_49
5_38
5_50
5_465_39
5_415_20
5_0
6_1
6_2
6_3
6_15
6_18
6_32
6_6
6_30
6_28
6_39
6_33
6_36
6_34
6_13
6_7
6_10
6_16
6_47
6_29
6_45
6_37
6_41
6_19
6_35
6_43
6_38
6_27
6_44
6_31
6_21
6_8
6_42
6_25
6_146_23
6_46
6_266_20
6_0
7_1
7_2
7_3
7_11
7_4
7_10
7_31
7_27
7_14
7_16
7_36
7_20
7_29
7_32
7_33
7_6
7_17
7_21
7_38
7_37
7_42
7_41
7_35
7_28
7_19
7_30
7_25
7_40
7_24
7_26
7_397_23
7_22
7_0
8_1
8_14
8_2
8_5
8_7
8_20
8_15
8_24
8_6
8_12
8_25
8_9
8_27
8_23
8_30
8_21
8_28
8_19
8_31
8_29
8_33
8_32
8_26
8_10
8_22
8_0
9_1
9_6
9_8
9_38
9_37
9_22
9_14
9_23
9_15
9_25
9_10
9_12
9_21
9_19
9_30
9_33
9_27
9_34
9_26
9_28
9_40
9_319_35
9_36
9_29
9_24
9_5
9_41
9_39
9_32
9_16
9_189_0
10_1
10_9
10_7
10_11
10_5
10_21
10_20
10_16
10_4
10_23
10_22
10_15
10_26
10_30
10_24
10_25
10_29
10_31
10_27
10_32
10_34
10_10
10_18
10_28
10_35
10_33
10_0
11_1
11_4
11_6
11_12
11_10
11_16
11_30
11_25
11_27
11_11
11_17
11_7
11_33
11_32
11_29
11_28
11_18
11_39
11_36
11_34
11_26
11_37
11_38
11_20
11_14
11_35
11_31
11_19
11_2
11_21
11_0
12_1
12_2
12_21
12_3
12_13
12_6
12_12
12_8
12_26
12_9
12_15
12_17
12_24
12_23
12_28
12_22
12_30
12_25
12_20
12_29
12_27
12_14
13_1
13_2
13_4
13_8
13_3
13_13
13_27
13_21
13_5
13_19
13_22
13_24
13_23
13_9
13_17
13_36
13_30
13_3413_26
13_32
13_28
13_31
13_14
13_18
13_33
13_2913_35
14_1
14_2
14_4
14_3
14_5
14_8
14_30
14_11
14_18
14_21
14_10
14_22
14_27
14_23
14_19
14_24
14_28
14_25
14_29
14_26
14_0
movie GBP.USD 2014movie EUR.USD 2014
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Step 2: lead-lag SVNs
determine groups
group lead-lag?
SNVs: p-values ofsg(t)→ sg′(t+1)
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Intertemporal interaction networks (FX LB)
Compute p-value of {sg(t),sg′(t+1)}
1
3
4[ −1 −> −1 ][ 1 −> −1 ][ 2 −> −1 ]
4
3 1
5[ −1 −> 1 ][ 1 −> 1 ]
1
3
4
5
[ 0 −> 0 ]
9
1430
32
1713
28
5
7
23
11
19
33
10
37
16
4
24
36
3
29
316
1
18
26
8
35
34
21
22
12
15
2527
[ −1 −> −1 ][ 0 −> −1 ][ 1 −> −1 ][ 2 −> −1 ]
7
23
25
14
28
32
96
151
4
5
16
17
18
2 3
810
29
13
11
20
3436
26
35
21
24
30
31
37
19
27
33[ −1 −> 0 ][ 0 −> 0 ][ 1 −> 0 ]
9
1
3
10
18
35
4
6
13
141627
31
32
36
37
7
332428
5
17
8
30
15
34
20
19
21
11
29
26
25
2
12[ −1 −> 1 ][ 0 −> 1 ][ 1 −> 1 ][ 2 −> 1 ]
33
25
35
13
24
15
10
5
11
32
19
6
8
[ −1 −> 2 ][ 1 −> 2 ][ 2 −> 2 ]
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
EUR.USD (LB)
9
1430
32
1713
28
5
7
23
11
19
33
10
37
16
4
24
36
3
29
316
1
18
26
8
35
34
21
22
12
15
2527
[ −1 −> −1 ][ 0 −> −1 ][ 1 −> −1 ][ 2 −> −1 ]
7
23
25
14
28
32
96
151
4
5
16
17
18
2 3
810
29
13
11
20
3436
26
35
21
24
30
31
37
19
27
33[ −1 −> 0 ][ 0 −> 0 ][ 1 −> 0 ]
9
1
3
10
18
35
4
6
13
141627
31
32
36
37
7
332428
5
17
8
30
15
34
20
19
21
11
29
26
25
2
12[ −1 −> 1 ][ 0 −> 1 ][ 1 −> 1 ][ 2 −> 1 ]
33
25
35
13
24
15
10
5
11
32
19
6
8
[ −1 −> 2 ][ 1 −> 2 ][ 2 −> 2 ]
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Origin of lead-lag
1 Explicit communication2 Delayed reaction
1 faster news2 same strategy, faster parameters
MA(CD) 12, 26, 5→ 10, 25, 4:3 Master in Finance + ...
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Lead-lag: number of links vs time
20
40
60
05−
2012
06−
2012
07−
2012
08−
2012
#Lin
ks
0
5
10
15
20
04−
2013
07−
2013
10−
2013
01−
2014
04−
2014
07−
2014
10−
2014
#Lin
ks
SQ LBEUR.USD
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Lead-lag between groups: inter-temporal networks
Goal: predict global flux sign (B−S) out of sample1 Sliding in-out sample periods2 Determine groups in-sample3 Calibrate machine learning in-sample4 Predict next SIGN
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Predictors+predictee
F1,t F2,t F1,t−1 F2,t−1 hour3 -2 NA 0 1
NA NA 3 -2 21 1 NA NA 3
sign (B−S)+1-1
????
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Machine learning: random forest
Random forest = collection of random trees
Classification tree: P predictors, T data points eachEach tree: bootstrap of data pointsEach node: cut according to criterion e.g. predictori > 2
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Prediction results
12 weeksin-sample
24 hoursout-of-sample
Predict sign offlux
Predictors:group actions{-1,0,1,NA}RandomForests
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Prediction results
12 weeksin-sample
24 hoursout-of-sample
Predict sign offlux
Predictors:group actions{-1,0,1,NA}RandomForests
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Prediction results: SQ
12 weeksin-sample
24 hoursout-of-sample
Predict sign offlux
Predictors:group actions{-1,0,1,NA}RandomForests
●●●
●●●●●●●●●●●●●●●●●
●●
●●
●●●●●●●●●●●●●●●●
●●●●●●
●
●●●●●●
●
●
●●●●●●●●●
●●
●●●●●●●●●
●●●●●●●●●
●●●●●●●●●●●●●●●●●●
●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●
●●●●●●●●●●●●●●●●●
●●●●●●●●●●●●●●●●●●●●●
●●●●●
●●●●●●●●●●●
●●●●
●●●●●●●●●●●●●●●●●●●●●●●●
●●●●●●●●●●●
●●
●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●
●●●●●●
●●●●●
●
●●
●●
●
●●●●●
●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●
●●
●
●●●●●●●●●●●●●●●●●●●
●●●●●●●
●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●
●●●●●
●●●●●●●●●●●●●●●●
●●●●●●
●●●●●●
●●●●●●●●●●●●●●●●
●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●
●●●●●●●●●
●
●
●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●
●●●●●●●●●●●●●●●●●●●●●●●●●●
●●●●●●●●●●●●●●●●
●●●●●●●●●●●●●●
●●●●●●●●●●●●●●●●●●●●●●●●●
●●●●●●●●●●●●●●●●●●●
●●●●●●●
●●●●●●●●●●●●●●●
●●●●●●●●●●●●
●●●●●●●●●●●
●●●●●●●●
●●●
●●●●●●●●●●●●●●●●●●●●●●●●●
●●●
●
●●●●●●●●●●●●●●
●●●●●●●●●●●●●●●●●●●●●●●●●●●
●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●
●●●●●●●●●●●●●●●●●●●●●
●●●●●●●●●●●●●●●
●●●●●●●●●●●●●●●●●●●●●
●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●
●●●●●●●●●●●●●●●●●●●●●●
●●●●●●●●●●●●●●●●●●●●●●
●●●●●●●●●●●●●●●●●●●●●●●●●
●●●●●●●●●●●●●
●●●●●●●●●●●●●●●●●●●●●●●
●
●●●●●●●●●●●●●●●●●
●●●●●●
●●●●●●●●●
●●●●
●●●●●●●●●●●●●●●●●●●●●
●●●●●●●●●●●●●●●●●●
●●●●●●●●●●●●●●●●●●●●●●
●●●●●●●●●●
●●●●●●●●●●●●●●●●●●●●●●●●
●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●
●●●●●●●●●●●
●●●●●●●●●●●●●●●●●●●●●●●●●●
●●●●●●
●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●
0 200 600 1000
0.0
0.5
1.0
1.5
2.0
2.5
Time (hours)
cum
sum
(stφ
t) in
bill
ions
● obliqueRFRFH0
2
H01
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Out-of-sample performance by hour
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Origins of performance peaks
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks
Summary
Method
Clustering→communities→ lead-lag→ machine learning→prediction
TodoFind algorithmic strategy of groups
Other fields
fund ownership
recommendations systems
...
D. Challet, M. Lallouache, R. Chicheportiche Predicting investment fluxes from implicit lead-lag investor networks