notebook.community

Edit and run 





In [1]:



    


import os, sys
sys.path.insert(0, os.path.abspath(os.path.join("..", "..")))

Prospective hotspotting





In [2]:



    


%matplotlib inline
import matplotlib.pyplot as plt
import open_cp.scripted
import open_cp.scripted.analysis as analysis



    











In [3]:



    


betas = analysis.hit_counts_to_beta("pro_grid_opt.csv")



    











In [4]:



    


import re
def label(key):
    m = re.search("Classic\((.*?)\)", key)
    return m.group(1)
label(list(betas)[0])



    


















    
Out[4]:








'sb=2, tb=6'

















In [5]:



    


fig, ax = plt.subplots(ncols=2, figsize=(16,7))
analysis.plot_betas(betas, ax[0])
ax[0].legend([])
analysis.plot_betas(betas, ax[1], plot_sds=False)
ax[1].legend([label(k) for k in betas])



    


















    
Out[5]:








<matplotlib.legend.Legend at 0x7f4f5e4d52e8>










    
























In [6]:



    


fig, ax = plt.subplots(ncols=2, figsize=(16,7))
analysis.plot_betas(betas, ax[0], range(1, 21))
ax[0].legend([])
analysis.plot_betas(betas, ax[1], range(1, 21), plot_sds=False)
ax[1].legend([label(k) for k in betas])



    


















    
Out[6]:








<matplotlib.legend.Legend at 0x7f4f5e3726a0>










    
























In [7]:



    


fig, ax = plt.subplots(figsize=(12,6))
r = analysis.plot_betas_means_against_max(betas, ax, range(1,21))



    


















    
























In [8]:



    


{k : all(x>0.85 for x in r[k]) for k in r}



    


















    
Out[8]:








{'ProHotspotProvider(Weight=Classic(sb=2, tb=6), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=2, tb=7), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=2, tb=8), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=2, tb=9), Distance=DiagsCircle, TimeUnit=24.0h)': True,
 'ProHotspotProvider(Weight=Classic(sb=3, tb=6), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=3, tb=7), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=3, tb=8), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=3, tb=9), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=4, tb=6), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=4, tb=7), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=4, tb=8), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=4, tb=9), Distance=DiagsCircle, TimeUnit=24.0h)': False}

















In [9]:



    


{k : sum(x>0.95 for x in r[k]) >= 16 for k in r}



    


















    
Out[9]:








{'ProHotspotProvider(Weight=Classic(sb=2, tb=6), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=2, tb=7), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=2, tb=8), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=2, tb=9), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=3, tb=6), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=3, tb=7), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=3, tb=8), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=3, tb=9), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=4, tb=6), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=4, tb=7), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=4, tb=8), Distance=DiagsCircle, TimeUnit=24.0h)': False,
 'ProHotspotProvider(Weight=Classic(sb=4, tb=9), Distance=DiagsCircle, TimeUnit=24.0h)': True}

















In [10]:



    


highlight = {"sb=2, tb=9", "sb=4, tb=9"}

fig, ax = plt.subplots(figsize=(10,6))
x = list(range(1, 21))
for k in r:
    if label(k) not in highlight:
        ax.plot(x, r[k], color="black", label="_")
    else:
        ax.plot(x, r[k], label=label(k))
ax.legend()
None

Continuous predictions





In [12]:



    


betas = analysis.hit_counts_to_beta("pro_opt.csv")



    











In [13]:



    


fig, ax = plt.subplots(ncols=2, figsize=(16,7))
analysis.plot_betas(betas, ax[0])
ax[0].legend([])
analysis.plot_betas(betas, ax[1], plot_sds=False)
ax[1].legend([label(k) for k in betas])



    


















    
Out[13]:








<matplotlib.legend.Legend at 0x7f4f60f0eeb8>










    
























In [14]:



    


fig, ax = plt.subplots(ncols=2, figsize=(16,7))
analysis.plot_betas(betas, ax[0], range(1, 21))
ax[0].legend([])
analysis.plot_betas(betas, ax[1], range(1, 21), plot_sds=False)
ax[1].legend([label(k) for k in betas])



    


















    
Out[14]:








<matplotlib.legend.Legend at 0x7f4f61143c18>










    
























In [15]:



    


fig, ax = plt.subplots(figsize=(12,6))
r = analysis.plot_betas_means_against_max(betas, ax, range(1,21))



    


















    
























In [19]:



    


{k : all(x>0.92 for x in r[k]) for k in r}



    


















    
Out[19]:








{'ProHotspotCtsProvider(Weight=Classic(sb=2, tb=6), DistanceUnit=150)': False,
 'ProHotspotCtsProvider(Weight=Classic(sb=2, tb=7), DistanceUnit=150)': False,
 'ProHotspotCtsProvider(Weight=Classic(sb=2, tb=8), DistanceUnit=150)': False,
 'ProHotspotCtsProvider(Weight=Classic(sb=2, tb=9), DistanceUnit=150)': False,
 'ProHotspotCtsProvider(Weight=Classic(sb=3, tb=6), DistanceUnit=150)': False,
 'ProHotspotCtsProvider(Weight=Classic(sb=3, tb=7), DistanceUnit=150)': False,
 'ProHotspotCtsProvider(Weight=Classic(sb=3, tb=8), DistanceUnit=150)': False,
 'ProHotspotCtsProvider(Weight=Classic(sb=3, tb=9), DistanceUnit=150)': False,
 'ProHotspotCtsProvider(Weight=Classic(sb=4, tb=6), DistanceUnit=150)': False,
 'ProHotspotCtsProvider(Weight=Classic(sb=4, tb=7), DistanceUnit=150)': True,
 'ProHotspotCtsProvider(Weight=Classic(sb=4, tb=8), DistanceUnit=150)': False,
 'ProHotspotCtsProvider(Weight=Classic(sb=4, tb=9), DistanceUnit=150)': False}

















In [22]:



    


{k : sum(x>0.96 for x in r[k]) >= 16 for k in r}



    


















    
Out[22]:








{'ProHotspotCtsProvider(Weight=Classic(sb=2, tb=6), DistanceUnit=150)': False,
 'ProHotspotCtsProvider(Weight=Classic(sb=2, tb=7), DistanceUnit=150)': False,
 'ProHotspotCtsProvider(Weight=Classic(sb=2, tb=8), DistanceUnit=150)': True,
 'ProHotspotCtsProvider(Weight=Classic(sb=2, tb=9), DistanceUnit=150)': True,
 'ProHotspotCtsProvider(Weight=Classic(sb=3, tb=6), DistanceUnit=150)': False,
 'ProHotspotCtsProvider(Weight=Classic(sb=3, tb=7), DistanceUnit=150)': False,
 'ProHotspotCtsProvider(Weight=Classic(sb=3, tb=8), DistanceUnit=150)': False,
 'ProHotspotCtsProvider(Weight=Classic(sb=3, tb=9), DistanceUnit=150)': True,
 'ProHotspotCtsProvider(Weight=Classic(sb=4, tb=6), DistanceUnit=150)': False,
 'ProHotspotCtsProvider(Weight=Classic(sb=4, tb=7), DistanceUnit=150)': False,
 'ProHotspotCtsProvider(Weight=Classic(sb=4, tb=8), DistanceUnit=150)': False,
 'ProHotspotCtsProvider(Weight=Classic(sb=4, tb=9), DistanceUnit=150)': False}

















In [25]:



    


highlight = {"sb=4, tb=7", "sb=3, tb=9"}

fig, ax = plt.subplots(figsize=(10,6))
x = list(range(1, 21))
for k in r:
    if label(k) not in highlight:
        ax.plot(x, r[k], color="black", label="_")
    else:
        ax.plot(x, r[k], label=label(k))
ax.legend()
None



    


















    
























In [ ]:

Content source: QuantCrimAtLeeds/PredictCode

Similar notebooks:

notebook.community | gallery | about