From 8e09ca6ca68c71bdab65525b529e2adfa281823c Mon Sep 17 00:00:00 2001
From: Ben Green <ben@SEASITs-MacBook-Pro.local>
Date: Wed, 1 Jul 2015 00:49:23 -0400
Subject: Got predict-victims running in parallel, drastically reducing the
 time for each test. Also changed how we get the rankings of infected
 individuals each day.

---
 R Scripts/generate-network.R | 40 +++++++++++++++++++++-------------------
 1 file changed, 21 insertions(+), 19 deletions(-)

(limited to 'R Scripts/generate-network.R')

diff --git a/R Scripts/generate-network.R b/R Scripts/generate-network.R
index 3b40969..dab81a4 100644
--- a/R Scripts/generate-network.R	
+++ b/R Scripts/generate-network.R	
@@ -3,20 +3,20 @@ setwd("~/Documents/Cascade Project/")
 source('criminal_cascades/R Scripts/temporal.R')
 source('criminal_cascades/R Scripts/structural.R')
 
-alpha = 1/100
-beta = 0.02
-delta = 0.15
+alpha = 1/10
+beta = 0.01
+delta = 0.25
 # lmbda = 1/10
 t_max = 1000
 
 N = 5000
 g = forest.fire.game(nodes=N, fw.prob=0.3, ambs=1, directed=F)
-plot(g, vertex.size=5, vertex.label=NA)
+plot(g, vertex.size=3, vertex.label=NA)
 
 V(g)$seed = runif(vcount(g))<beta
 seeds = which(V(g)$seed)
 V(g)$vic = V(g)$seed
-V(g)$vic.day[V(g)$seed] = sample(1:t_max, sum(V(g)$seed))
+V(g)$vic.day[V(g)$seed] = 1#sample(1:t_max, sum(V(g)$seed))
 V(g)$spawn.date = 0
 V(g)$infector = NA
 
@@ -28,17 +28,19 @@ for (day in 1:t_max){
     dists = as.numeric(shortest.paths(g,vic,neighbors))
     infected = neighbors[which(runif(length(neighbors))<structural(delta, dists))]
     infected = setdiff(infected,seeds) # don't try to infect seeds
-    inf.days = day + ceiling(alpha*rexp(length(infected),alpha))
+    inf.days = day + ceiling(rexp(length(infected),alpha))
+    realized = ((inf.days <= V(g)$vic.day[infected]) %in% c(NA,T)) & (inf.days<=t_max)
+    infected = infected[realized]
     V(g)$vic[infected] = TRUE
-    infects = (inf.days <= V(g)$vic.day[infected]) %in% c(NA,T)
-    V(g)$vic.day[infected[infects]] = inf.days[infects]
-    V(g)$infector[infected[infects]] = vic
+    V(g)$vic.day[infected] = inf.days[realized]
+    V(g)$infector[infected] = vic
   }
 }
 
 vic_ids = which(V(g)$vic)
+print(length(vic_ids))
 cols = rep('lightblue',N); cols[V(g)$vic]='red'; cols[V(g)$seed]='darkred'
-plot(g, vertex.size=5, vertex.label=NA, vertex.color=cols)
+plot(g, vertex.size=3, vertex.label=NA, vertex.color=cols)
 
 ##### generate dag_dat
 dag_dat_test = data.frame(matrix(nrow=1,ncol=10))
@@ -68,12 +70,12 @@ rownames(dag_dat_test) = NULL
 
 write.csv(dag_dat_test, file='Results/dag_dat_test.csv')
 
-##### analyze performance of recovery algorithm
-recovered = read.csv('Results/infectors.csv',header=F,col.names=c('victim','infector'))
-recovered = recovered[order(recovered$victim),]
-infectors = cbind(setdiff(vic_ids,seeds),
-                  V(g)$infector[setdiff(vic_ids,seeds)],
-                  recovered$infector[recovered$victim %in% setdiff(vic_ids,seeds)])
-mean(infectors[,2]==infectors[,3])
-
-dag_dat_test[dag_dat_test$to==4984,]
\ No newline at end of file
+##### analyze performance of recovery algorithm ------
+# recovered = read.csv('Results/infectors.csv',header=F,col.names=c('victim','infector'))
+# recovered = recovered[order(recovered$victim),]
+# infectors = cbind(setdiff(vic_ids,seeds),
+#                   V(g)$infector[setdiff(vic_ids,seeds)],
+#                   recovered$infector[recovered$victim %in% setdiff(vic_ids,seeds)])
+# mean(infectors[,2]==infectors[,3])
+# 
+# dag_dat_test[dag_dat_test$to==4984,]
-- 
cgit v1.2.3-70-g09d2


From 110069d77815a3d62e3526f18b2a34fb79beff1e Mon Sep 17 00:00:00 2001
From: Ben Green <ben@SEASITs-MacBook-Pro.local>
Date: Thu, 2 Jul 2015 23:10:54 -0400
Subject: just some minor tests with ml

---
 R Scripts/generate-network.R                       |  16 +-
 experiments/build/temp.macosx-10.6-x86_64-2.7/ml.o | Bin 306196 -> 306268 bytes
 experiments/build_network.py                       |   2 +-
 experiments/ml.c                                   | 180 +++++++++----------
 experiments/ml.pyx                                 |  11 +-
 experiments/ml.so                                  | Bin 119284 -> 119276 bytes
 experiments/out.log                                | 200 ++++++++++-----------
 experiments/process.py                             |   4 +-
 8 files changed, 207 insertions(+), 206 deletions(-)

(limited to 'R Scripts/generate-network.R')

diff --git a/R Scripts/generate-network.R b/R Scripts/generate-network.R
index dab81a4..db7012d 100644
--- a/R Scripts/generate-network.R	
+++ b/R Scripts/generate-network.R	
@@ -3,9 +3,9 @@ setwd("~/Documents/Cascade Project/")
 source('criminal_cascades/R Scripts/temporal.R')
 source('criminal_cascades/R Scripts/structural.R')
 
-alpha = 1/10
-beta = 0.01
-delta = 0.25
+alpha = 0.1
+beta = 0.25
+delta = 0.1
 # lmbda = 1/10
 t_max = 1000
 
@@ -16,7 +16,7 @@ plot(g, vertex.size=3, vertex.label=NA)
 V(g)$seed = runif(vcount(g))<beta
 seeds = which(V(g)$seed)
 V(g)$vic = V(g)$seed
-V(g)$vic.day[V(g)$seed] = 1#sample(1:t_max, sum(V(g)$seed))
+V(g)$vic.day[V(g)$seed] = sample(1:t_max, sum(V(g)$seed), replace=T) #1 if testing ml2
 V(g)$spawn.date = 0
 V(g)$infector = NA
 
@@ -31,9 +31,11 @@ for (day in 1:t_max){
     inf.days = day + ceiling(rexp(length(infected),alpha))
     realized = ((inf.days <= V(g)$vic.day[infected]) %in% c(NA,T)) & (inf.days<=t_max)
     infected = infected[realized]
-    V(g)$vic[infected] = TRUE
-    V(g)$vic.day[infected] = inf.days[realized]
-    V(g)$infector[infected] = vic
+    if(sum(realized)){
+      V(g)$vic[infected] = TRUE
+      V(g)$vic.day[infected] = inf.days[realized]
+      V(g)$infector[infected] = vic
+    }
   }
 }
 
diff --git a/experiments/build/temp.macosx-10.6-x86_64-2.7/ml.o b/experiments/build/temp.macosx-10.6-x86_64-2.7/ml.o
index 8667998..8cabba1 100644
Binary files a/experiments/build/temp.macosx-10.6-x86_64-2.7/ml.o and b/experiments/build/temp.macosx-10.6-x86_64-2.7/ml.o differ
diff --git a/experiments/build_network.py b/experiments/build_network.py
index a481a90..c58d5cf 100644
--- a/experiments/build_network.py
+++ b/experiments/build_network.py
@@ -8,7 +8,7 @@ def build_network(filename):
     victims = {}
     non_victims = {}
     age = 0
-    t_max = 3012
+    t_max = 1000
     with open(filename) as fh:
         reader = DictReader(fh)
         for row in reader:
diff --git a/experiments/ml.c b/experiments/ml.c
index 92e313f..4b9c2ef 100644
--- a/experiments/ml.c
+++ b/experiments/ml.c
@@ -1332,7 +1332,7 @@ static PyObject *__pyx_n_s_zeros;
 static PyObject *__pyx_float_1_;
 static PyObject *__pyx_float__01;
 static PyObject *__pyx_float_0_01;
-static PyObject *__pyx_int_148152;
+static PyObject *__pyx_int_5000;
 static PyObject *__pyx_tuple_;
 static PyObject *__pyx_tuple__2;
 static PyObject *__pyx_tuple__3;
@@ -1363,7 +1363,7 @@ static __pyx_t_2ml_DTYPE_t __pyx_f_2ml_weight_success(int __pyx_v_dist, int __py
  *     """weight for successful infection, exponential time model"""
  *     cdef DTYPE_t structural, temporal, result
  *     structural = delta ** dist             # <<<<<<<<<<<<<<
- *     temporal = log(exp(alpha)-1.) - alpha*dt/7.
+ *     temporal = log(exp(alpha)-1.) - alpha*dt/1.
  *     result = log(structural) + temporal
  */
   __pyx_v_structural = pow(__pyx_v_delta, ((__pyx_t_2ml_DTYPE_t)__pyx_v_dist));
@@ -1371,15 +1371,15 @@ static __pyx_t_2ml_DTYPE_t __pyx_f_2ml_weight_success(int __pyx_v_dist, int __py
   /* "ml.pyx":14
  *     cdef DTYPE_t structural, temporal, result
  *     structural = delta ** dist
- *     temporal = log(exp(alpha)-1.) - alpha*dt/7.             # <<<<<<<<<<<<<<
+ *     temporal = log(exp(alpha)-1.) - alpha*dt/1.             # <<<<<<<<<<<<<<
  *     result = log(structural) + temporal
  *     return result
  */
-  __pyx_v_temporal = (log((exp(__pyx_v_alpha) - 1.)) - ((__pyx_v_alpha * __pyx_v_dt) / 7.));
+  __pyx_v_temporal = (log((exp(__pyx_v_alpha) - 1.)) - ((__pyx_v_alpha * __pyx_v_dt) / 1.));
 
   /* "ml.pyx":15
  *     structural = delta ** dist
- *     temporal = log(exp(alpha)-1.) - alpha*dt/7.
+ *     temporal = log(exp(alpha)-1.) - alpha*dt/1.
  *     result = log(structural) + temporal             # <<<<<<<<<<<<<<
  *     return result
  * 
@@ -1387,7 +1387,7 @@ static __pyx_t_2ml_DTYPE_t __pyx_f_2ml_weight_success(int __pyx_v_dist, int __py
   __pyx_v_result = (log(__pyx_v_structural) + __pyx_v_temporal);
 
   /* "ml.pyx":16
- *     temporal = log(exp(alpha)-1.) - alpha*dt/7.
+ *     temporal = log(exp(alpha)-1.) - alpha*dt/1.
  *     result = log(structural) + temporal
  *     return result             # <<<<<<<<<<<<<<
  * 
@@ -1430,7 +1430,7 @@ static __pyx_t_2ml_DTYPE_t __pyx_f_2ml_weight_failure(int __pyx_v_dist, int __py
  *     """weight for failed infection, exponential time model"""
  *     cdef DTYPE_t structural, temporal, result
  *     structural = delta ** dist             # <<<<<<<<<<<<<<
- *     temporal = exp(-alpha * dt/7.)
+ *     temporal = exp(-alpha * dt/1.)
  *     result = log(1. - structural + structural * temporal)
  */
   __pyx_v_structural = pow(__pyx_v_delta, ((__pyx_t_2ml_DTYPE_t)__pyx_v_dist));
@@ -1438,15 +1438,15 @@ static __pyx_t_2ml_DTYPE_t __pyx_f_2ml_weight_failure(int __pyx_v_dist, int __py
   /* "ml.pyx":23
  *     cdef DTYPE_t structural, temporal, result
  *     structural = delta ** dist
- *     temporal = exp(-alpha * dt/7.)             # <<<<<<<<<<<<<<
+ *     temporal = exp(-alpha * dt/1.)             # <<<<<<<<<<<<<<
  *     result = log(1. - structural + structural * temporal)
  *     return result
  */
-  __pyx_v_temporal = exp((((-__pyx_v_alpha) * __pyx_v_dt) / 7.));
+  __pyx_v_temporal = exp((((-__pyx_v_alpha) * __pyx_v_dt) / 1.));
 
   /* "ml.pyx":24
  *     structural = delta ** dist
- *     temporal = exp(-alpha * dt/7.)
+ *     temporal = exp(-alpha * dt/1.)
  *     result = log(1. - structural + structural * temporal)             # <<<<<<<<<<<<<<
  *     return result
  * 
@@ -1454,7 +1454,7 @@ static __pyx_t_2ml_DTYPE_t __pyx_f_2ml_weight_failure(int __pyx_v_dist, int __py
   __pyx_v_result = log(((1. - __pyx_v_structural) + (__pyx_v_structural * __pyx_v_temporal)));
 
   /* "ml.pyx":25
- *     temporal = exp(-alpha * dt/7.)
+ *     temporal = exp(-alpha * dt/1.)
  *     result = log(1. - structural + structural * temporal)
  *     return result             # <<<<<<<<<<<<<<
  * 
@@ -1701,7 +1701,7 @@ static PyObject *__pyx_pf_2ml_ml(CYTHON_UNUSED PyObject *__pyx_self, PyObject *_
  *         DTYPE_t beta, ll, beta_add, max_beta, max_beta_add
  *         list parents, failures, successes
  *     n_roots, n_victims = len(root_victims), len(victims)             # <<<<<<<<<<<<<<
- *     n_nodes = 148152
+ *     n_nodes = 5000
  *     cdef:
  */
   if (unlikely(__pyx_v_root_victims == Py_None)) {
@@ -1720,15 +1720,15 @@ static PyObject *__pyx_pf_2ml_ml(CYTHON_UNUSED PyObject *__pyx_self, PyObject *_
   /* "ml.pyx":34
  *         list parents, failures, successes
  *     n_roots, n_victims = len(root_victims), len(victims)
- *     n_nodes = 148152             # <<<<<<<<<<<<<<
+ *     n_nodes = 5000             # <<<<<<<<<<<<<<
  *     cdef:
  *         np.ndarray[DTYPE_t] probs = np.zeros(n_victims, dtype=DTYPE)
  */
-  __Pyx_INCREF(__pyx_int_148152);
-  __pyx_v_n_nodes = __pyx_int_148152;
+  __Pyx_INCREF(__pyx_int_5000);
+  __pyx_v_n_nodes = __pyx_int_5000;
 
   /* "ml.pyx":36
- *     n_nodes = 148152
+ *     n_nodes = 5000
  *     cdef:
  *         np.ndarray[DTYPE_t] probs = np.zeros(n_victims, dtype=DTYPE)             # <<<<<<<<<<<<<<
  *         np.ndarray[DTYPE_t] probs_fail = np.zeros(n_victims, dtype=DTYPE)
@@ -2858,7 +2858,7 @@ static PyObject *__pyx_pf_2ml_ml(CYTHON_UNUSED PyObject *__pyx_self, PyObject *_
  *     ll =  probs_fail.sum() # add probability that all edges to all victims fail
  *     ll += probs_nv.sum() # add probability that all edges to non_victims fail             # <<<<<<<<<<<<<<
  * 
- *     # print 'probs', probs
+ *     max_beta_add = float('-inf')
  */
   __pyx_t_4 = PyFloat_FromDouble(__pyx_v_ll); if (unlikely(!__pyx_t_4)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 75; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
   __Pyx_GOTREF(__pyx_t_4);
@@ -2890,38 +2890,38 @@ static PyObject *__pyx_pf_2ml_ml(CYTHON_UNUSED PyObject *__pyx_self, PyObject *_
   __Pyx_DECREF(__pyx_t_5); __pyx_t_5 = 0;
   __pyx_v_ll = __pyx_t_27;
 
-  /* "ml.pyx":78
+  /* "ml.pyx":77
+ *     ll += probs_nv.sum() # add probability that all edges to non_victims fail
  * 
- *     # print 'probs', probs
  *     max_beta_add = float('-inf')             # <<<<<<<<<<<<<<
  *     # iterate over all victim nodes to find the optimal threshold
  *     for beta in np.arange(0.01, 1., .01):
  */
-  __pyx_t_29 = __Pyx_PyObject_AsDouble(__pyx_kp_s_inf); if (unlikely(__pyx_t_29 == ((double)-1) && PyErr_Occurred())) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 78; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+  __pyx_t_29 = __Pyx_PyObject_AsDouble(__pyx_kp_s_inf); if (unlikely(__pyx_t_29 == ((double)-1) && PyErr_Occurred())) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 77; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
   __pyx_v_max_beta_add = __pyx_t_29;
 
-  /* "ml.pyx":80
+  /* "ml.pyx":79
  *     max_beta_add = float('-inf')
  *     # iterate over all victim nodes to find the optimal threshold
  *     for beta in np.arange(0.01, 1., .01):             # <<<<<<<<<<<<<<
- *         thresh = log(beta/(3012.*(1.-beta)))
+ *         thresh = log(beta/(1000.*(1.-beta)))
  *         seeds = probs<thresh
  */
-  __pyx_t_5 = __Pyx_GetModuleGlobalName(__pyx_n_s_np); if (unlikely(!__pyx_t_5)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 80; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+  __pyx_t_5 = __Pyx_GetModuleGlobalName(__pyx_n_s_np); if (unlikely(!__pyx_t_5)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 79; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
   __Pyx_GOTREF(__pyx_t_5);
-  __pyx_t_6 = __Pyx_PyObject_GetAttrStr(__pyx_t_5, __pyx_n_s_arange); if (unlikely(!__pyx_t_6)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 80; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+  __pyx_t_6 = __Pyx_PyObject_GetAttrStr(__pyx_t_5, __pyx_n_s_arange); if (unlikely(!__pyx_t_6)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 79; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
   __Pyx_GOTREF(__pyx_t_6);
   __Pyx_DECREF(__pyx_t_5); __pyx_t_5 = 0;
-  __pyx_t_5 = __Pyx_PyObject_Call(__pyx_t_6, __pyx_tuple_, NULL); if (unlikely(!__pyx_t_5)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 80; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+  __pyx_t_5 = __Pyx_PyObject_Call(__pyx_t_6, __pyx_tuple_, NULL); if (unlikely(!__pyx_t_5)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 79; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
   __Pyx_GOTREF(__pyx_t_5);
   __Pyx_DECREF(__pyx_t_6); __pyx_t_6 = 0;
   if (likely(PyList_CheckExact(__pyx_t_5)) || PyTuple_CheckExact(__pyx_t_5)) {
     __pyx_t_6 = __pyx_t_5; __Pyx_INCREF(__pyx_t_6); __pyx_t_2 = 0;
     __pyx_t_36 = NULL;
   } else {
-    __pyx_t_2 = -1; __pyx_t_6 = PyObject_GetIter(__pyx_t_5); if (unlikely(!__pyx_t_6)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 80; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_2 = -1; __pyx_t_6 = PyObject_GetIter(__pyx_t_5); if (unlikely(!__pyx_t_6)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 79; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_GOTREF(__pyx_t_6);
-    __pyx_t_36 = Py_TYPE(__pyx_t_6)->tp_iternext; if (unlikely(!__pyx_t_36)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 80; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_36 = Py_TYPE(__pyx_t_6)->tp_iternext; if (unlikely(!__pyx_t_36)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 79; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
   }
   __Pyx_DECREF(__pyx_t_5); __pyx_t_5 = 0;
   for (;;) {
@@ -2929,16 +2929,16 @@ static PyObject *__pyx_pf_2ml_ml(CYTHON_UNUSED PyObject *__pyx_self, PyObject *_
       if (likely(PyList_CheckExact(__pyx_t_6))) {
         if (__pyx_t_2 >= PyList_GET_SIZE(__pyx_t_6)) break;
         #if CYTHON_COMPILING_IN_CPYTHON
-        __pyx_t_5 = PyList_GET_ITEM(__pyx_t_6, __pyx_t_2); __Pyx_INCREF(__pyx_t_5); __pyx_t_2++; if (unlikely(0 < 0)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 80; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+        __pyx_t_5 = PyList_GET_ITEM(__pyx_t_6, __pyx_t_2); __Pyx_INCREF(__pyx_t_5); __pyx_t_2++; if (unlikely(0 < 0)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 79; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
         #else
-        __pyx_t_5 = PySequence_ITEM(__pyx_t_6, __pyx_t_2); __pyx_t_2++; if (unlikely(!__pyx_t_5)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 80; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+        __pyx_t_5 = PySequence_ITEM(__pyx_t_6, __pyx_t_2); __pyx_t_2++; if (unlikely(!__pyx_t_5)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 79; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
         #endif
       } else {
         if (__pyx_t_2 >= PyTuple_GET_SIZE(__pyx_t_6)) break;
         #if CYTHON_COMPILING_IN_CPYTHON
-        __pyx_t_5 = PyTuple_GET_ITEM(__pyx_t_6, __pyx_t_2); __Pyx_INCREF(__pyx_t_5); __pyx_t_2++; if (unlikely(0 < 0)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 80; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+        __pyx_t_5 = PyTuple_GET_ITEM(__pyx_t_6, __pyx_t_2); __Pyx_INCREF(__pyx_t_5); __pyx_t_2++; if (unlikely(0 < 0)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 79; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
         #else
-        __pyx_t_5 = PySequence_ITEM(__pyx_t_6, __pyx_t_2); __pyx_t_2++; if (unlikely(!__pyx_t_5)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 80; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+        __pyx_t_5 = PySequence_ITEM(__pyx_t_6, __pyx_t_2); __pyx_t_2++; if (unlikely(!__pyx_t_5)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 79; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
         #endif
       }
     } else {
@@ -2947,79 +2947,79 @@ static PyObject *__pyx_pf_2ml_ml(CYTHON_UNUSED PyObject *__pyx_self, PyObject *_
         PyObject* exc_type = PyErr_Occurred();
         if (exc_type) {
           if (likely(exc_type == PyExc_StopIteration || PyErr_GivenExceptionMatches(exc_type, PyExc_StopIteration))) PyErr_Clear();
-          else {__pyx_filename = __pyx_f[0]; __pyx_lineno = 80; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+          else {__pyx_filename = __pyx_f[0]; __pyx_lineno = 79; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
         }
         break;
       }
       __Pyx_GOTREF(__pyx_t_5);
     }
-    __pyx_t_27 = __pyx_PyFloat_AsDouble(__pyx_t_5); if (unlikely((__pyx_t_27 == (npy_double)-1) && PyErr_Occurred())) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 80; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_27 = __pyx_PyFloat_AsDouble(__pyx_t_5); if (unlikely((__pyx_t_27 == (npy_double)-1) && PyErr_Occurred())) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 79; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_DECREF(__pyx_t_5); __pyx_t_5 = 0;
     __pyx_v_beta = __pyx_t_27;
 
-    /* "ml.pyx":81
+    /* "ml.pyx":80
  *     # iterate over all victim nodes to find the optimal threshold
  *     for beta in np.arange(0.01, 1., .01):
- *         thresh = log(beta/(3012.*(1.-beta)))             # <<<<<<<<<<<<<<
+ *         thresh = log(beta/(1000.*(1.-beta)))             # <<<<<<<<<<<<<<
  *         seeds = probs<thresh
  *         non_seeds = probs>=thresh
  */
-    __pyx_v_thresh = log((__pyx_v_beta / (3012. * (1. - __pyx_v_beta))));
+    __pyx_v_thresh = log((__pyx_v_beta / (1000. * (1. - __pyx_v_beta))));
 
-    /* "ml.pyx":82
+    /* "ml.pyx":81
  *     for beta in np.arange(0.01, 1., .01):
- *         thresh = log(beta/(3012.*(1.-beta)))
+ *         thresh = log(beta/(1000.*(1.-beta)))
  *         seeds = probs<thresh             # <<<<<<<<<<<<<<
  *         non_seeds = probs>=thresh
  *         roots = n_roots + sum(seeds)
  */
-    __pyx_t_5 = PyFloat_FromDouble(__pyx_v_thresh); if (unlikely(!__pyx_t_5)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 82; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_5 = PyFloat_FromDouble(__pyx_v_thresh); if (unlikely(!__pyx_t_5)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 81; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_GOTREF(__pyx_t_5);
-    __pyx_t_4 = PyObject_RichCompare(((PyObject *)__pyx_v_probs), __pyx_t_5, Py_LT); __Pyx_XGOTREF(__pyx_t_4); if (unlikely(!__pyx_t_4)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 82; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_4 = PyObject_RichCompare(((PyObject *)__pyx_v_probs), __pyx_t_5, Py_LT); __Pyx_XGOTREF(__pyx_t_4); if (unlikely(!__pyx_t_4)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 81; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_DECREF(__pyx_t_5); __pyx_t_5 = 0;
     __Pyx_XDECREF_SET(__pyx_v_seeds, __pyx_t_4);
     __pyx_t_4 = 0;
 
-    /* "ml.pyx":83
- *         thresh = log(beta/(3012.*(1.-beta)))
+    /* "ml.pyx":82
+ *         thresh = log(beta/(1000.*(1.-beta)))
  *         seeds = probs<thresh
  *         non_seeds = probs>=thresh             # <<<<<<<<<<<<<<
  *         roots = n_roots + sum(seeds)
  * 
  */
-    __pyx_t_4 = PyFloat_FromDouble(__pyx_v_thresh); if (unlikely(!__pyx_t_4)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 83; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_4 = PyFloat_FromDouble(__pyx_v_thresh); if (unlikely(!__pyx_t_4)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 82; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_GOTREF(__pyx_t_4);
-    __pyx_t_5 = PyObject_RichCompare(((PyObject *)__pyx_v_probs), __pyx_t_4, Py_GE); __Pyx_XGOTREF(__pyx_t_5); if (unlikely(!__pyx_t_5)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 83; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_5 = PyObject_RichCompare(((PyObject *)__pyx_v_probs), __pyx_t_4, Py_GE); __Pyx_XGOTREF(__pyx_t_5); if (unlikely(!__pyx_t_5)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 82; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_DECREF(__pyx_t_4); __pyx_t_4 = 0;
     __Pyx_XDECREF_SET(__pyx_v_non_seeds, __pyx_t_5);
     __pyx_t_5 = 0;
 
-    /* "ml.pyx":84
+    /* "ml.pyx":83
  *         seeds = probs<thresh
  *         non_seeds = probs>=thresh
  *         roots = n_roots + sum(seeds)             # <<<<<<<<<<<<<<
  * 
  *         beta_add = 0.
  */
-    __pyx_t_5 = __Pyx_PyInt_From_int(__pyx_v_n_roots); if (unlikely(!__pyx_t_5)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 84; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_5 = __Pyx_PyInt_From_int(__pyx_v_n_roots); if (unlikely(!__pyx_t_5)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 83; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_GOTREF(__pyx_t_5);
-    __pyx_t_4 = PyTuple_New(1); if (unlikely(!__pyx_t_4)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 84; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_4 = PyTuple_New(1); if (unlikely(!__pyx_t_4)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 83; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_GOTREF(__pyx_t_4);
     __Pyx_INCREF(__pyx_v_seeds);
     PyTuple_SET_ITEM(__pyx_t_4, 0, __pyx_v_seeds);
     __Pyx_GIVEREF(__pyx_v_seeds);
-    __pyx_t_3 = __Pyx_PyObject_Call(__pyx_builtin_sum, __pyx_t_4, NULL); if (unlikely(!__pyx_t_3)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 84; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_3 = __Pyx_PyObject_Call(__pyx_builtin_sum, __pyx_t_4, NULL); if (unlikely(!__pyx_t_3)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 83; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_GOTREF(__pyx_t_3);
     __Pyx_DECREF(__pyx_t_4); __pyx_t_4 = 0;
-    __pyx_t_4 = PyNumber_Add(__pyx_t_5, __pyx_t_3); if (unlikely(!__pyx_t_4)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 84; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_4 = PyNumber_Add(__pyx_t_5, __pyx_t_3); if (unlikely(!__pyx_t_4)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 83; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_GOTREF(__pyx_t_4);
     __Pyx_DECREF(__pyx_t_5); __pyx_t_5 = 0;
     __Pyx_DECREF(__pyx_t_3); __pyx_t_3 = 0;
-    __pyx_t_12 = __Pyx_PyInt_As_int(__pyx_t_4); if (unlikely((__pyx_t_12 == (int)-1) && PyErr_Occurred())) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 84; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_12 = __Pyx_PyInt_As_int(__pyx_t_4); if (unlikely((__pyx_t_12 == (int)-1) && PyErr_Occurred())) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 83; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_DECREF(__pyx_t_4); __pyx_t_4 = 0;
     __pyx_v_roots = __pyx_t_12;
 
-    /* "ml.pyx":86
+    /* "ml.pyx":85
  *         roots = n_roots + sum(seeds)
  * 
  *         beta_add = 0.             # <<<<<<<<<<<<<<
@@ -3028,18 +3028,18 @@ static PyObject *__pyx_pf_2ml_ml(CYTHON_UNUSED PyObject *__pyx_self, PyObject *_
  */
     __pyx_v_beta_add = 0.;
 
-    /* "ml.pyx":88
+    /* "ml.pyx":87
  *         beta_add = 0.
  *         # add probability for realized edges and subtract probability these edges fail
  *         beta_add += (probs[non_seeds]).sum()             # <<<<<<<<<<<<<<
  *         # add probability for the seeds and non-seeds
- *         beta_add += roots * log(beta/3012.) + (n_nodes-roots) * log(1. - beta)
+ *         beta_add += roots * log(beta/1000.) + (n_nodes-roots) * log(1. - beta)
  */
-    __pyx_t_4 = PyFloat_FromDouble(__pyx_v_beta_add); if (unlikely(!__pyx_t_4)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 88; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_4 = PyFloat_FromDouble(__pyx_v_beta_add); if (unlikely(!__pyx_t_4)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 87; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_GOTREF(__pyx_t_4);
-    __pyx_t_5 = PyObject_GetItem(((PyObject *)__pyx_v_probs), __pyx_v_non_seeds); if (unlikely(__pyx_t_5 == NULL)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 88; __pyx_clineno = __LINE__; goto __pyx_L1_error;};
+    __pyx_t_5 = PyObject_GetItem(((PyObject *)__pyx_v_probs), __pyx_v_non_seeds); if (unlikely(__pyx_t_5 == NULL)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 87; __pyx_clineno = __LINE__; goto __pyx_L1_error;};
     __Pyx_GOTREF(__pyx_t_5);
-    __pyx_t_21 = __Pyx_PyObject_GetAttrStr(__pyx_t_5, __pyx_n_s_sum); if (unlikely(!__pyx_t_21)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 88; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_21 = __Pyx_PyObject_GetAttrStr(__pyx_t_5, __pyx_n_s_sum); if (unlikely(!__pyx_t_21)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 87; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_GOTREF(__pyx_t_21);
     __Pyx_DECREF(__pyx_t_5); __pyx_t_5 = 0;
     __pyx_t_5 = NULL;
@@ -3053,57 +3053,57 @@ static PyObject *__pyx_pf_2ml_ml(CYTHON_UNUSED PyObject *__pyx_self, PyObject *_
       }
     }
     if (__pyx_t_5) {
-      __pyx_t_3 = __Pyx_PyObject_CallOneArg(__pyx_t_21, __pyx_t_5); if (unlikely(!__pyx_t_3)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 88; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+      __pyx_t_3 = __Pyx_PyObject_CallOneArg(__pyx_t_21, __pyx_t_5); if (unlikely(!__pyx_t_3)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 87; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
       __Pyx_DECREF(__pyx_t_5); __pyx_t_5 = 0;
     } else {
-      __pyx_t_3 = __Pyx_PyObject_CallNoArg(__pyx_t_21); if (unlikely(!__pyx_t_3)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 88; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+      __pyx_t_3 = __Pyx_PyObject_CallNoArg(__pyx_t_21); if (unlikely(!__pyx_t_3)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 87; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     }
     __Pyx_GOTREF(__pyx_t_3);
     __Pyx_DECREF(__pyx_t_21); __pyx_t_21 = 0;
-    __pyx_t_21 = PyNumber_InPlaceAdd(__pyx_t_4, __pyx_t_3); if (unlikely(!__pyx_t_21)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 88; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_21 = PyNumber_InPlaceAdd(__pyx_t_4, __pyx_t_3); if (unlikely(!__pyx_t_21)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 87; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_GOTREF(__pyx_t_21);
     __Pyx_DECREF(__pyx_t_4); __pyx_t_4 = 0;
     __Pyx_DECREF(__pyx_t_3); __pyx_t_3 = 0;
-    __pyx_t_27 = __pyx_PyFloat_AsDouble(__pyx_t_21); if (unlikely((__pyx_t_27 == (npy_double)-1) && PyErr_Occurred())) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 88; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_27 = __pyx_PyFloat_AsDouble(__pyx_t_21); if (unlikely((__pyx_t_27 == (npy_double)-1) && PyErr_Occurred())) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 87; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_DECREF(__pyx_t_21); __pyx_t_21 = 0;
     __pyx_v_beta_add = __pyx_t_27;
 
-    /* "ml.pyx":90
+    /* "ml.pyx":89
  *         beta_add += (probs[non_seeds]).sum()
  *         # add probability for the seeds and non-seeds
- *         beta_add += roots * log(beta/3012.) + (n_nodes-roots) * log(1. - beta)             # <<<<<<<<<<<<<<
+ *         beta_add += roots * log(beta/1000.) + (n_nodes-roots) * log(1. - beta)             # <<<<<<<<<<<<<<
  * 
  *         if beta_add > max_beta_add:
  */
-    __pyx_t_21 = PyFloat_FromDouble(__pyx_v_beta_add); if (unlikely(!__pyx_t_21)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 90; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_21 = PyFloat_FromDouble(__pyx_v_beta_add); if (unlikely(!__pyx_t_21)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 89; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_GOTREF(__pyx_t_21);
-    __pyx_t_3 = PyFloat_FromDouble((__pyx_v_roots * log((__pyx_v_beta / 3012.)))); if (unlikely(!__pyx_t_3)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 90; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_3 = PyFloat_FromDouble((__pyx_v_roots * log((__pyx_v_beta / 1000.)))); if (unlikely(!__pyx_t_3)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 89; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_GOTREF(__pyx_t_3);
-    __pyx_t_4 = __Pyx_PyInt_From_int(__pyx_v_roots); if (unlikely(!__pyx_t_4)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 90; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_4 = __Pyx_PyInt_From_int(__pyx_v_roots); if (unlikely(!__pyx_t_4)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 89; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_GOTREF(__pyx_t_4);
-    __pyx_t_5 = PyNumber_Subtract(__pyx_v_n_nodes, __pyx_t_4); if (unlikely(!__pyx_t_5)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 90; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_5 = PyNumber_Subtract(__pyx_v_n_nodes, __pyx_t_4); if (unlikely(!__pyx_t_5)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 89; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_GOTREF(__pyx_t_5);
     __Pyx_DECREF(__pyx_t_4); __pyx_t_4 = 0;
-    __pyx_t_4 = PyFloat_FromDouble(log((1. - __pyx_v_beta))); if (unlikely(!__pyx_t_4)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 90; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_4 = PyFloat_FromDouble(log((1. - __pyx_v_beta))); if (unlikely(!__pyx_t_4)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 89; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_GOTREF(__pyx_t_4);
-    __pyx_t_20 = PyNumber_Multiply(__pyx_t_5, __pyx_t_4); if (unlikely(!__pyx_t_20)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 90; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_20 = PyNumber_Multiply(__pyx_t_5, __pyx_t_4); if (unlikely(!__pyx_t_20)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 89; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_GOTREF(__pyx_t_20);
     __Pyx_DECREF(__pyx_t_5); __pyx_t_5 = 0;
     __Pyx_DECREF(__pyx_t_4); __pyx_t_4 = 0;
-    __pyx_t_4 = PyNumber_Add(__pyx_t_3, __pyx_t_20); if (unlikely(!__pyx_t_4)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 90; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_4 = PyNumber_Add(__pyx_t_3, __pyx_t_20); if (unlikely(!__pyx_t_4)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 89; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_GOTREF(__pyx_t_4);
     __Pyx_DECREF(__pyx_t_3); __pyx_t_3 = 0;
     __Pyx_DECREF(__pyx_t_20); __pyx_t_20 = 0;
-    __pyx_t_20 = PyNumber_InPlaceAdd(__pyx_t_21, __pyx_t_4); if (unlikely(!__pyx_t_20)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 90; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_20 = PyNumber_InPlaceAdd(__pyx_t_21, __pyx_t_4); if (unlikely(!__pyx_t_20)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 89; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_GOTREF(__pyx_t_20);
     __Pyx_DECREF(__pyx_t_21); __pyx_t_21 = 0;
     __Pyx_DECREF(__pyx_t_4); __pyx_t_4 = 0;
-    __pyx_t_27 = __pyx_PyFloat_AsDouble(__pyx_t_20); if (unlikely((__pyx_t_27 == (npy_double)-1) && PyErr_Occurred())) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 90; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+    __pyx_t_27 = __pyx_PyFloat_AsDouble(__pyx_t_20); if (unlikely((__pyx_t_27 == (npy_double)-1) && PyErr_Occurred())) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 89; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
     __Pyx_DECREF(__pyx_t_20); __pyx_t_20 = 0;
     __pyx_v_beta_add = __pyx_t_27;
 
-    /* "ml.pyx":92
- *         beta_add += roots * log(beta/3012.) + (n_nodes-roots) * log(1. - beta)
+    /* "ml.pyx":91
+ *         beta_add += roots * log(beta/1000.) + (n_nodes-roots) * log(1. - beta)
  * 
  *         if beta_add > max_beta_add:             # <<<<<<<<<<<<<<
  *             max_beta = beta
@@ -3112,7 +3112,7 @@ static PyObject *__pyx_pf_2ml_ml(CYTHON_UNUSED PyObject *__pyx_self, PyObject *_
     __pyx_t_31 = ((__pyx_v_beta_add > __pyx_v_max_beta_add) != 0);
     if (__pyx_t_31) {
 
-      /* "ml.pyx":93
+      /* "ml.pyx":92
  * 
  *         if beta_add > max_beta_add:
  *             max_beta = beta             # <<<<<<<<<<<<<<
@@ -3121,7 +3121,7 @@ static PyObject *__pyx_pf_2ml_ml(CYTHON_UNUSED PyObject *__pyx_self, PyObject *_
  */
       __pyx_v_max_beta = __pyx_v_beta;
 
-      /* "ml.pyx":94
+      /* "ml.pyx":93
  *         if beta_add > max_beta_add:
  *             max_beta = beta
  *             max_roots = roots             # <<<<<<<<<<<<<<
@@ -3130,7 +3130,7 @@ static PyObject *__pyx_pf_2ml_ml(CYTHON_UNUSED PyObject *__pyx_self, PyObject *_
  */
       __pyx_v_max_roots = __pyx_v_roots;
 
-      /* "ml.pyx":95
+      /* "ml.pyx":94
  *             max_beta = beta
  *             max_roots = roots
  *             max_beta_add = beta_add             # <<<<<<<<<<<<<<
@@ -3142,17 +3142,17 @@ static PyObject *__pyx_pf_2ml_ml(CYTHON_UNUSED PyObject *__pyx_self, PyObject *_
     }
     __pyx_L32:;
 
-    /* "ml.pyx":80
+    /* "ml.pyx":79
  *     max_beta_add = float('-inf')
  *     # iterate over all victim nodes to find the optimal threshold
  *     for beta in np.arange(0.01, 1., .01):             # <<<<<<<<<<<<<<
- *         thresh = log(beta/(3012.*(1.-beta)))
+ *         thresh = log(beta/(1000.*(1.-beta)))
  *         seeds = probs<thresh
  */
   }
   __Pyx_DECREF(__pyx_t_6); __pyx_t_6 = 0;
 
-  /* "ml.pyx":98
+  /* "ml.pyx":97
  *             # print 'beta:', max_beta, 'add:', max_beta_add, 'roots:', max_roots
  * 
  *     ll += max_beta_add             # <<<<<<<<<<<<<<
@@ -3161,7 +3161,7 @@ static PyObject *__pyx_pf_2ml_ml(CYTHON_UNUSED PyObject *__pyx_self, PyObject *_
  */
   __pyx_v_ll = (__pyx_v_ll + __pyx_v_max_beta_add);
 
-  /* "ml.pyx":99
+  /* "ml.pyx":98
  * 
  *     ll += max_beta_add
  *     roots = max_roots             # <<<<<<<<<<<<<<
@@ -3170,7 +3170,7 @@ static PyObject *__pyx_pf_2ml_ml(CYTHON_UNUSED PyObject *__pyx_self, PyObject *_
  */
   __pyx_v_roots = __pyx_v_max_roots;
 
-  /* "ml.pyx":100
+  /* "ml.pyx":99
  *     ll += max_beta_add
  *     roots = max_roots
  *     beta = max_beta             # <<<<<<<<<<<<<<
@@ -3179,19 +3179,19 @@ static PyObject *__pyx_pf_2ml_ml(CYTHON_UNUSED PyObject *__pyx_self, PyObject *_
  */
   __pyx_v_beta = __pyx_v_max_beta;
 
-  /* "ml.pyx":102
+  /* "ml.pyx":101
  *     beta = max_beta
  *     # print n_nodes, n_roots, n_victims, max_i, roots
  *     return (beta, roots, ll)             # <<<<<<<<<<<<<<
  */
   __Pyx_XDECREF(__pyx_r);
-  __pyx_t_6 = PyFloat_FromDouble(__pyx_v_beta); if (unlikely(!__pyx_t_6)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 102; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+  __pyx_t_6 = PyFloat_FromDouble(__pyx_v_beta); if (unlikely(!__pyx_t_6)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 101; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
   __Pyx_GOTREF(__pyx_t_6);
-  __pyx_t_20 = __Pyx_PyInt_From_int(__pyx_v_roots); if (unlikely(!__pyx_t_20)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 102; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+  __pyx_t_20 = __Pyx_PyInt_From_int(__pyx_v_roots); if (unlikely(!__pyx_t_20)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 101; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
   __Pyx_GOTREF(__pyx_t_20);
-  __pyx_t_4 = PyFloat_FromDouble(__pyx_v_ll); if (unlikely(!__pyx_t_4)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 102; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+  __pyx_t_4 = PyFloat_FromDouble(__pyx_v_ll); if (unlikely(!__pyx_t_4)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 101; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
   __Pyx_GOTREF(__pyx_t_4);
-  __pyx_t_21 = PyTuple_New(3); if (unlikely(!__pyx_t_21)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 102; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+  __pyx_t_21 = PyTuple_New(3); if (unlikely(!__pyx_t_21)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 101; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
   __Pyx_GOTREF(__pyx_t_21);
   PyTuple_SET_ITEM(__pyx_t_21, 0, __pyx_t_6);
   __Pyx_GIVEREF(__pyx_t_6);
@@ -5376,14 +5376,14 @@ static int __Pyx_InitCachedConstants(void) {
   __Pyx_RefNannyDeclarations
   __Pyx_RefNannySetupContext("__Pyx_InitCachedConstants", 0);
 
-  /* "ml.pyx":80
+  /* "ml.pyx":79
  *     max_beta_add = float('-inf')
  *     # iterate over all victim nodes to find the optimal threshold
  *     for beta in np.arange(0.01, 1., .01):             # <<<<<<<<<<<<<<
- *         thresh = log(beta/(3012.*(1.-beta)))
+ *         thresh = log(beta/(1000.*(1.-beta)))
  *         seeds = probs<thresh
  */
-  __pyx_tuple_ = PyTuple_Pack(3, __pyx_float_0_01, __pyx_float_1_, __pyx_float__01); if (unlikely(!__pyx_tuple_)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 80; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+  __pyx_tuple_ = PyTuple_Pack(3, __pyx_float_0_01, __pyx_float_1_, __pyx_float__01); if (unlikely(!__pyx_tuple_)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 79; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
   __Pyx_GOTREF(__pyx_tuple_);
   __Pyx_GIVEREF(__pyx_tuple_);
 
@@ -5476,7 +5476,7 @@ static int __Pyx_InitGlobals(void) {
   __pyx_float_1_ = PyFloat_FromDouble(1.); if (unlikely(!__pyx_float_1_)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 1; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
   __pyx_float__01 = PyFloat_FromDouble(.01); if (unlikely(!__pyx_float__01)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 1; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
   __pyx_float_0_01 = PyFloat_FromDouble(0.01); if (unlikely(!__pyx_float_0_01)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 1; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
-  __pyx_int_148152 = PyInt_FromLong(148152L); if (unlikely(!__pyx_int_148152)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 1; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
+  __pyx_int_5000 = PyInt_FromLong(5000); if (unlikely(!__pyx_int_5000)) {__pyx_filename = __pyx_f[0]; __pyx_lineno = 1; __pyx_clineno = __LINE__; goto __pyx_L1_error;}
   return 0;
   __pyx_L1_error:;
   return -1;
diff --git a/experiments/ml.pyx b/experiments/ml.pyx
index 858207f..0d72c44 100644
--- a/experiments/ml.pyx
+++ b/experiments/ml.pyx
@@ -11,7 +11,7 @@ cdef DTYPE_t weight_success(int dist, int dt, DTYPE_t alpha, DTYPE_t delta,
     """weight for successful infection, exponential time model"""
     cdef DTYPE_t structural, temporal, result
     structural = delta ** dist
-    temporal = log(exp(alpha)-1.) - alpha*dt/7.
+    temporal = log(exp(alpha)-1.) - alpha*dt/1.
     result = log(structural) + temporal
     return result
 
@@ -20,7 +20,7 @@ cdef DTYPE_t weight_failure(int dist, int dt, DTYPE_t alpha, DTYPE_t delta,
     """weight for failed infection, exponential time model"""
     cdef DTYPE_t structural, temporal, result
     structural = delta ** dist
-    temporal = exp(-alpha * dt/7.)
+    temporal = exp(-alpha * dt/1.)
     result = log(1. - structural + structural * temporal)
     return result
 
@@ -31,7 +31,7 @@ def ml(dict root_victims, dict victims, dict non_victims, DTYPE_t age,
         DTYPE_t beta, ll, beta_add, max_beta, max_beta_add
         list parents, failures, successes
     n_roots, n_victims = len(root_victims), len(victims)
-    n_nodes = 148152
+    n_nodes = 5000
     cdef:
         np.ndarray[DTYPE_t] probs = np.zeros(n_victims, dtype=DTYPE)
         np.ndarray[DTYPE_t] probs_fail = np.zeros(n_victims, dtype=DTYPE)
@@ -74,11 +74,10 @@ def ml(dict root_victims, dict victims, dict non_victims, DTYPE_t age,
     ll =  probs_fail.sum() # add probability that all edges to all victims fail
     ll += probs_nv.sum() # add probability that all edges to non_victims fail
 
-    # print 'probs', probs
     max_beta_add = float('-inf')
     # iterate over all victim nodes to find the optimal threshold
     for beta in np.arange(0.01, 1., .01):
-        thresh = log(beta/(3012.*(1.-beta)))
+        thresh = log(beta/(1000.*(1.-beta)))
         seeds = probs<thresh
         non_seeds = probs>=thresh
         roots = n_roots + sum(seeds)
@@ -87,7 +86,7 @@ def ml(dict root_victims, dict victims, dict non_victims, DTYPE_t age,
         # add probability for realized edges and subtract probability these edges fail
         beta_add += (probs[non_seeds]).sum()
         # add probability for the seeds and non-seeds
-        beta_add += roots * log(beta/3012.) + (n_nodes-roots) * log(1. - beta)
+        beta_add += roots * log(beta/1000.) + (n_nodes-roots) * log(1. - beta)
 
         if beta_add > max_beta_add:
             max_beta = beta
diff --git a/experiments/ml.so b/experiments/ml.so
index a08d635..d4cd7b4 100755
Binary files a/experiments/ml.so and b/experiments/ml.so differ
diff --git a/experiments/out.log b/experiments/out.log
index a718511..895f69f 100644
--- a/experiments/out.log
+++ b/experiments/out.log
@@ -1,100 +1,100 @@
-0.001	0.001	0.08	11270	-130174.021949
-0.001	0.0016681005372	0.08	11270	-130186.81792
-0.001	0.00278255940221	0.08	11270	-130208.510741
-0.001	0.00464158883361	0.08	11270	-130245.688267
-0.001	0.00774263682681	0.08	11270	-130310.57411
-0.001	0.0129154966501	0.08	11270	-130427.308991
-0.001	0.0215443469003	0.08	11270	-130648.06397
-0.001	0.035938136638	0.04	6478	-128816.657559
-0.001	0.0599484250319	0.04	6478	-127394.665041
-0.001	0.1	0.04	6478	-127688.933813
-0.0016681005372	0.001	0.08	11270	-130184.321487
-0.0016681005372	0.0016681005372	0.08	11270	-130204.10481
-0.0016681005372	0.00278255940221	0.08	11270	-130237.646155
-0.0016681005372	0.00464158883361	0.08	11270	-130295.138019
-0.0016681005372	0.00774263682681	0.08	11270	-130395.500445
-0.0016681005372	0.0129154966501	0.08	11270	-130576.118953
-0.0016681005372	0.0215443469003	0.04	6478	-128846.541805
-0.0016681005372	0.035938136638	0.04	6478	-127087.253323
-0.0016681005372	0.0599484250319	0.04	6478	-126230.808822
-0.0016681005372	0.1	0.02	2792	-126557.181556
-0.00278255940221	0.001	0.08	11270	-130198.331202
-0.00278255940221	0.0016681005372	0.08	11270	-130227.620554
-0.00278255940221	0.00278255940221	0.08	11270	-130277.284253
-0.00278255940221	0.00464158883361	0.08	11270	-130362.425847
-0.00278255940221	0.00774263682681	0.08	11270	-130511.096608
-0.00278255940221	0.0129154966501	0.04	6499	-129059.422383
-0.00278255940221	0.0215443469003	0.04	6478	-127109.015024
-0.00278255940221	0.035938136638	0.04	6478	-125684.241338
-0.00278255940221	0.0599484250319	0.04	6478	-125600.024485
-0.00278255940221	0.1	0.02	2792	-126287.212393
-0.00464158883361	0.001	0.08	11270	-130215.294173
-0.00464158883361	0.0016681005372	0.08	11270	-130256.095445
-0.00464158883361	0.00278255940221	0.08	11270	-130325.28705
-0.00464158883361	0.00464158883361	0.08	11270	-130443.928528
-0.00464158883361	0.00774263682681	0.05	7027	-129395.352373
-0.00464158883361	0.0129154966501	0.04	6516	-127431.591365
-0.00464158883361	0.0215443469003	0.04	6478	-125683.037727
-0.00464158883361	0.035938136638	0.04	6478	-124664.01619
-0.00464158883361	0.0599484250319	0.02	2914	-124502.496739
-0.00464158883361	0.1	0.02	2792	-127134.604186
-0.00774263682681	0.001	0.08	11270	-130232.773853
-0.00774263682681	0.0016681005372	0.08	11270	-130285.439066
-0.00774263682681	0.00278255940221	0.08	11270	-130374.758088
-0.00774263682681	0.00464158883361	0.05	7652	-129744.991978
-0.00774263682681	0.00774263682681	0.05	6995	-127969.97982
-0.00774263682681	0.0129154966501	0.05	6677	-126168.108707
-0.00774263682681	0.0215443469003	0.04	6513	-124653.173909
-0.00774263682681	0.035938136638	0.04	6478	-124060.374554
-0.00774263682681	0.0599484250319	0.02	2937	-123459.631205
-0.00774263682681	0.1	0.01	1199	-128091.177099
-0.0129154966501	0.001	0.08	11270	-130247.637572
-0.0129154966501	0.0016681005372	0.08	11270	-130310.391126
-0.0129154966501	0.00278255940221	0.06	8952	-129978.711147
-0.0129154966501	0.00464158883361	0.05	7622	-128580.851845
-0.0129154966501	0.00774263682681	0.05	7183	-126905.034076
-0.0129154966501	0.0129154966501	0.05	6902	-125304.331411
-0.0129154966501	0.0215443469003	0.05	6712	-124093.056726
-0.0129154966501	0.035938136638	0.03	4220	-123607.130582
-0.0129154966501	0.0599484250319	0.02	3041	-122971.205801
-0.0129154966501	0.1	0.01	1210	-129142.326165
-0.0215443469003	0.001	0.08	11270	-130258.217698
-0.0215443469003	0.0016681005372	0.07	9957	-130081.814439
-0.0215443469003	0.00278255940221	0.06	8662	-129124.841865
-0.0215443469003	0.00464158883361	0.05	7864	-127781.895187
-0.0215443469003	0.00774263682681	0.05	7487	-126287.652103
-0.0215443469003	0.0129154966501	0.05	7248	-124916.428004
-0.0215443469003	0.0215443469003	0.05	7058	-124002.594007
-0.0215443469003	0.035938136638	0.03	4060	-122986.296269
-0.0215443469003	0.0599484250319	0.02	3225	-122943.105886
-0.0215443469003	0.1	0.01	1280	-129952.279246
-0.035938136638	0.001	0.07	10368	-130108.933188
-0.035938136638	0.0016681005372	0.06	9401	-129491.738149
-0.035938136638	0.00278255940221	0.06	8856	-128521.580165
-0.035938136638	0.00464158883361	0.06	8372	-127371.393158
-0.035938136638	0.00774263682681	0.05	7935	-126147.240183
-0.035938136638	0.0129154966501	0.05	7682	-125046.741534
-0.035938136638	0.0215443469003	0.05	7476	-124429.198394
-0.035938136638	0.035938136638	0.03	4307	-122878.240073
-0.035938136638	0.0599484250319	0.02	2939	-123316.885543
-0.035938136638	0.1	0.01	1389	-130593.383741
-0.0599484250319	0.001	0.07	10079	-129709.888924
-0.0599484250319	0.0016681005372	0.06	9545	-129062.475372
-0.0599484250319	0.00278255940221	0.06	9190	-128212.556391
-0.0599484250319	0.00464158883361	0.06	8868	-127288.891344
-0.0599484250319	0.00774263682681	0.06	8559	-126362.833809
-0.0599484250319	0.0129154966501	0.06	8304	-125604.769419
-0.0599484250319	0.0215443469003	0.04	6187	-124909.433209
-0.0599484250319	0.035938136638	0.03	4809	-123357.362127
-0.0599484250319	0.0599484250319	0.02	2564	-123821.010849
-0.0599484250319	0.1	0.01	1603	-131308.788018
-0.1	0.001	0.07	10122	-129407.183852
-0.1	0.0016681005372	0.07	9858	-128826.387421
-0.1	0.00278255940221	0.06	9551	-128164.459489
-0.1	0.00464158883361	0.06	9323	-127449.154213
-0.1	0.00774263682681	0.06	9138	-126798.726378
-0.1	0.0129154966501	0.06	8941	-126366.966378
-0.1	0.0215443469003	0.05	6820	-125456.586416
-0.1	0.035938136638	0.04	5608	-124150.372812
-0.1	0.0599484250319	0.02	2771	-124505.31667
-0.1	0.1	0.01	1909	-132157.286741
+0.1	0.01	0.32	1587	-16246.2835303
+0.1	0.0129154966501	0.3	1522	-16193.1560593
+0.1	0.016681005372	0.3	1482	-16136.7007139
+0.1	0.0215443469003	0.29	1445	-16083.927662
+0.1	0.0278255940221	0.28	1418	-16045.4325535
+0.1	0.035938136638	0.28	1399	-16033.3471288
+0.1	0.0464158883361	0.28	1381	-16065.8045855
+0.1	0.0599484250319	0.27	1365	-16174.1785652
+0.1	0.0774263682681	0.23	1166	-16354.0210508
+0.1	0.1	0.22	1094	-16710.6318304
+0.113646366639	0.01	0.32	1587	-16223.0544334
+0.113646366639	0.0129154966501	0.3	1522	-16169.9167648
+0.113646366639	0.016681005372	0.3	1496	-16114.7613028
+0.113646366639	0.0215443469003	0.29	1453	-16064.997114
+0.113646366639	0.0278255940221	0.29	1436	-16029.4224738
+0.113646366639	0.035938136638	0.28	1413	-16020.1691105
+0.113646366639	0.0464158883361	0.28	1399	-16057.1566663
+0.113646366639	0.0599484250319	0.26	1321	-16167.2369119
+0.113646366639	0.0774263682681	0.23	1173	-16335.8221545
+0.113646366639	0.1	0.22	1108	-16694.4584299
+0.129154966501	0.01	0.32	1587	-16201.5712361
+0.129154966501	0.0129154966501	0.31	1537	-16149.1696536
+0.129154966501	0.016681005372	0.3	1496	-16095.9372617
+0.129154966501	0.0215443469003	0.29	1464	-16049.6817227
+0.129154966501	0.0278255940221	0.29	1445	-16016.7348561
+0.129154966501	0.035938136638	0.29	1426	-16011.4090845
+0.129154966501	0.0464158883361	0.28	1407	-16051.4875514
+0.129154966501	0.0599484250319	0.26	1288	-16157.5499839
+0.129154966501	0.0774263682681	0.24	1183	-16321.8627529
+0.129154966501	0.1	0.22	1115	-16682.3833134
+0.146779926762	0.01	0.32	1587	-16182.0774435
+0.146779926762	0.0129154966501	0.31	1537	-16131.2048239
+0.146779926762	0.016681005372	0.3	1507	-16080.1708094
+0.146779926762	0.0215443469003	0.3	1482	-16037.2306303
+0.146779926762	0.0278255940221	0.29	1453	-16008.0169599
+0.146779926762	0.035938136638	0.29	1436	-16006.0071619
+0.146779926762	0.0464158883361	0.28	1418	-16050.0144027
+0.146779926762	0.0599484250319	0.26	1281	-16148.6892961
+0.146779926762	0.0774263682681	0.24	1187	-16311.3545786
+0.146779926762	0.1	0.23	1138	-16673.8963578
+0.16681005372	0.01	0.32	1587	-16164.849071
+0.16681005372	0.0129154966501	0.31	1560	-16116.449014
+0.16681005372	0.016681005372	0.3	1522	-16067.8766242
+0.16681005372	0.0215443469003	0.3	1496	-16027.3638618
+0.16681005372	0.0278255940221	0.29	1464	-16003.2426018
+0.16681005372	0.035938136638	0.29	1453	-16004.4084252
+0.16681005372	0.0464158883361	0.29	1436	-16052.4652191
+0.16681005372	0.0599484250319	0.26	1289	-16141.4404164
+0.16681005372	0.0774263682681	0.24	1196	-16303.5459122
+0.16681005372	0.1	0.23	1145	-16670.0806528
+0.189573565241	0.01	0.32	1613	-16149.7026241
+0.189573565241	0.0129154966501	0.31	1560	-16103.5965222
+0.189573565241	0.016681005372	0.31	1537	-16058.5451224
+0.189573565241	0.0215443469003	0.3	1507	-16020.9935104
+0.189573565241	0.0278255940221	0.3	1496	-16000.5729402
+0.189573565241	0.035938136638	0.29	1464	-16006.9538317
+0.189573565241	0.0464158883361	0.29	1453	-16058.7205642
+0.189573565241	0.0599484250319	0.25	1260	-16136.4254637
+0.189573565241	0.0774263682681	0.24	1206	-16299.8411955
+0.189573565241	0.1	0.23	1157	-16671.3131735
+0.215443469003	0.01	0.32	1613	-16136.2887288
+0.215443469003	0.0129154966501	0.32	1587	-16093.9635334
+0.215443469003	0.016681005372	0.31	1537	-16052.071471
+0.215443469003	0.0215443469003	0.3	1522	-16018.1621521
+0.215443469003	0.0278255940221	0.3	1507	-16000.7844822
+0.215443469003	0.035938136638	0.3	1482	-16012.1618981
+0.215443469003	0.0464158883361	0.28	1399	-16066.3079647
+0.215443469003	0.0599484250319	0.25	1272	-16136.5860246
+0.215443469003	0.0774263682681	0.24	1212	-16301.153347
+0.215443469003	0.1	0.23	1173	-16676.745682
+0.244843674682	0.01	0.32	1613	-16125.5232284
+0.244843674682	0.0129154966501	0.32	1587	-16085.6169883
+0.244843674682	0.016681005372	0.31	1560	-16048.1741694
+0.244843674682	0.0215443469003	0.31	1537	-16018.2052077
+0.244843674682	0.0278255940221	0.3	1522	-16004.8785245
+0.244843674682	0.035938136638	0.3	1507	-16019.6055296
+0.244843674682	0.0464158883361	0.28	1415	-16072.4792069
+0.244843674682	0.0599484250319	0.26	1278	-16139.4702448
+0.244843674682	0.0774263682681	0.25	1237	-16306.1951749
+0.244843674682	0.1	0.24	1195	-16688.1452238
+0.278255940221	0.01	0.33	1637	-16116.7453334
+0.278255940221	0.0129154966501	0.32	1613	-16080.0365035
+0.278255940221	0.016681005372	0.32	1587	-16046.9682889
+0.278255940221	0.0215443469003	0.31	1560	-16021.5867637
+0.278255940221	0.0278255940221	0.31	1537	-16011.7866661
+0.278255940221	0.035938136638	0.3	1522	-16030.8034204
+0.278255940221	0.0464158883361	0.27	1367	-16078.1951393
+0.278255940221	0.0599484250319	0.26	1298	-16144.7176332
+0.278255940221	0.0774263682681	0.25	1249	-16315.5245594
+0.278255940221	0.1	0.24	1207	-16702.3027221
+0.316227766017	0.01	0.33	1637	-16109.5865483
+0.316227766017	0.0129154966501	0.32	1613	-16075.7276647
+0.316227766017	0.016681005372	0.32	1587	-16046.7383729
+0.316227766017	0.0215443469003	0.31	1560	-16026.6264576
+0.316227766017	0.0278255940221	0.31	1537	-16022.2115914
+0.316227766017	0.035938136638	0.31	1537	-16044.7445633
+0.316227766017	0.0464158883361	0.27	1374	-16084.9299944
+0.316227766017	0.0599484250319	0.26	1308	-16152.3121983
+0.316227766017	0.0774263682681	0.25	1260	-16327.1609197
+0.316227766017	0.1	0.25	1235	-16718.4366844
diff --git a/experiments/process.py b/experiments/process.py
index a0c9b98..6c71934 100644
--- a/experiments/process.py
+++ b/experiments/process.py
@@ -16,9 +16,9 @@ if __name__ == "__main__":
     root_victims, victims, non_victims, age = load(open(sys.argv[1]))
     
     # alphas = np.arange(.02, .1, .01)  # parameter of the time component
-    alphas = np.logspace(-3,-1,num=10)
+    alphas = np.logspace(-1,-.5,num=10)
     # deltas = np.arange(.02, .1, .01)  # parameter of the structural component
-    deltas = np.logspace(-3,-1,num=10)
+    deltas = np.logspace(-2,-1,num=10)
     with open("out.log", "w") as fh:
         for alpha, delta in product(alphas, deltas):
             beta, roots, ll = ml(root_victims, victims, non_victims, age, alpha, delta)
-- 
cgit v1.2.3-70-g09d2