Peter B commited on
Commit
b5956de
1 Parent(s): 3b2813e

final model update

Browse files
Files changed (2) hide show
  1. README.md +85 -85
  2. pytorch_model.bin +1 -1
README.md CHANGED
@@ -49,7 +49,7 @@ The model in this repository is trained on all available data. Because of this,
49
 
50
  | Metric | precision | recall | f1-score |
51
  | --------- | --------- | ------ | -------- |
52
- | macro avg | 0.811 | 0.788 | 0.795 |
53
 
54
 
55
  *Note*: These are the average of the values *per fold*, so *macro avg* is the average of the macro average of all categories per fold.
@@ -58,90 +58,90 @@ The model in this repository is trained on all available data. Because of this,
58
 
59
  | Category | precision | recall | f1-score | support |
60
  | ------------------------------------------------------ | --------- | ------ | -------- | ------- |
61
- | AGGRAVATED ASSAULT | 0.95 | 0.957 | 0.954 | 4085 |
62
- | ARMED ROBBERY | 0.958 | 0.957 | 0.957 | 1021 |
63
- | ARSON | 0.951 | 0.953 | 0.952 | 344 |
64
- | ASSAULTING PUBLIC OFFICER | 0.921 | 0.905 | 0.913 | 588 |
65
- | AUTO THEFT | 0.962 | 0.963 | 0.963 | 1660 |
66
- | BLACKMAIL/EXTORTION/INTIMIDATION | 0.868 | 0.878 | 0.873 | 627 |
67
- | BRIBERY AND CONFLICT OF INTEREST | 0.772 | 0.821 | 0.795 | 216 |
68
- | BURGLARY | 0.982 | 0.98 | 0.981 | 2214 |
69
- | CHILD ABUSE | 0.793 | 0.776 | 0.784 | 139 |
70
- | COCAINE OR CRACK VIOLATION OFFENSE UNSPECIFIED | 0.868 | 0.865 | 0.866 | 47 |
71
- | COMMERCIALIZED VICE | 0.811 | 0.777 | 0.793 | 666 |
72
- | CONTEMPT OF COURT | 0.983 | 0.987 | 0.985 | 2952 |
73
- | CONTRIBUTING TO DELINQUENCY OF A MINOR | 0.454 | 0.379 | 0.396 | 50 |
74
- | CONTROLLED SUBSTANCE - OFFENSE UNSPECIFIED | 0.839 | 0.78 | 0.808 | 280 |
75
  | COUNTERFEITING (FEDERAL ONLY) | 0 | 0 | 0 | 2 |
76
- | DESTRUCTION OF PROPERTY | 0.967 | 0.97 | 0.969 | 2560 |
77
- | DRIVING UNDER INFLUENCE - DRUGS | 0.641 | 0.616 | 0.625 | 34 |
78
- | DRIVING UNDER THE INFLUENCE | 0.944 | 0.95 | 0.947 | 2195 |
79
- | DRIVING WHILE INTOXICATED | 0.99 | 0.978 | 0.984 | 2391 |
80
- | DRUG OFFENSES - VIOLATION/DRUG UNSPECIFIED | 0.905 | 0.911 | 0.908 | 3100 |
81
- | DRUNKENNESS/VAGRANCY/DISORDERLY CONDUCT | 0.853 | 0.862 | 0.857 | 380 |
82
- | EMBEZZLEMENT | 0.86 | 0.762 | 0.808 | 100 |
83
  | EMBEZZLEMENT (FEDERAL ONLY) | 0 | 0 | 0 | 1 |
84
- | ESCAPE FROM CUSTODY | 0.989 | 0.99 | 0.99 | 4035 |
85
- | FAMILY RELATED OFFENSES | 0.744 | 0.768 | 0.756 | 442 |
86
- | FELONY - UNSPECIFIED | 0.665 | 0.762 | 0.709 | 122 |
87
- | FLIGHT TO AVOID PROSECUTION | 0.43 | 0.434 | 0.431 | 38 |
88
- | FORCIBLE SODOMY | 0.773 | 0.837 | 0.802 | 76 |
89
  | FORGERY (FEDERAL ONLY) | 0 | 0 | 0 | 2 |
90
- | FORGERY/FRAUD | 0.909 | 0.928 | 0.918 | 4687 |
91
  | FRAUD (FEDERAL ONLY) | 0 | 0 | 0 | 2 |
92
- | GRAND LARCENY - THEFT OVER $200 | 0.959 | 0.972 | 0.966 | 2412 |
93
- | HABITUAL OFFENDER | 0.748 | 0.656 | 0.695 | 53 |
94
- | HEROIN VIOLATION - OFFENSE UNSPECIFIED | 0.877 | 0.777 | 0.82 | 24 |
95
- | HIT AND RUN DRIVING | 0.927 | 0.933 | 0.93 | 303 |
96
- | HIT/RUN DRIVING - PROPERTY DAMAGE | 0.929 | 0.924 | 0.926 | 362 |
97
- | IMMIGRATION VIOLATIONS | 0.778 | 0.616 | 0.681 | 19 |
98
- | INVASION OF PRIVACY | 0.925 | 0.923 | 0.924 | 1235 |
99
- | JUVENILE OFFENSES | 0.9 | 0.871 | 0.883 | 144 |
100
- | KIDNAPPING | 0.926 | 0.929 | 0.927 | 553 |
101
- | LARCENY/THEFT - VALUE UNKNOWN | 0.953 | 0.946 | 0.95 | 3175 |
102
- | LEWD ACT WITH CHILDREN | 0.786 | 0.846 | 0.814 | 596 |
103
- | LIQUOR LAW VIOLATIONS | 0.731 | 0.762 | 0.746 | 214 |
104
- | MANSLAUGHTER - NON-VEHICULAR | 0.661 | 0.803 | 0.725 | 139 |
105
- | MANSLAUGHTER - VEHICULAR | 0.763 | 0.854 | 0.803 | 117 |
106
- | MARIJUANA/HASHISH VIOLATION - OFFENSE UNSPECIFIED | 0.778 | 0.675 | 0.718 | 62 |
107
- | MISDEMEANOR UNSPECIFIED | 0.616 | 0.256 | 0.357 | 57 |
108
- | MORALS/DECENCY - OFFENSE | 0.759 | 0.763 | 0.761 | 412 |
109
- | MURDER | 0.965 | 0.922 | 0.943 | 621 |
110
- | OBSTRUCTION - LAW ENFORCEMENT | 0.945 | 0.947 | 0.946 | 4220 |
111
- | OFFENSES AGAINST COURTS, LEGISLATURES, AND COMMISSIONS | 0.882 | 0.897 | 0.889 | 1965 |
112
- | PAROLE VIOLATION | 0.968 | 0.949 | 0.958 | 946 |
113
- | PETTY LARCENY - THEFT UNDER $200 | 0.987 | 0.768 | 0.864 | 139 |
114
- | POSSESSION/USE - COCAINE OR CRACK | 0.9 | 0.928 | 0.913 | 68 |
115
- | POSSESSION/USE - DRUG UNSPECIFIED | 0.618 | 0.561 | 0.586 | 189 |
116
- | POSSESSION/USE - HEROIN | 0.917 | 0.839 | 0.876 | 25 |
117
- | POSSESSION/USE - MARIJUANA/HASHISH | 0.975 | 0.973 | 0.974 | 556 |
118
- | POSSESSION/USE - OTHER CONTROLLED SUBSTANCES | 0.976 | 0.965 | 0.97 | 3271 |
119
- | PROBATION VIOLATION | 0.958 | 0.956 | 0.957 | 1158 |
120
- | PROPERTY OFFENSES - OTHER | 0.892 | 0.863 | 0.878 | 446 |
121
- | PUBLIC ORDER OFFENSES - OTHER | 0.706 | 0.717 | 0.711 | 1871 |
122
  | RACKETEERING/EXTORTION (FEDERAL ONLY) | 0 | 0 | 0 | 2 |
123
- | RAPE - FORCE | 0.841 | 0.871 | 0.856 | 641 |
124
- | RAPE - STATUTORY - NO FORCE | 0.714 | 0.551 | 0.619 | 140 |
125
- | REGULATORY OFFENSES (FEDERAL ONLY) | 0.8 | 0.558 | 0.657 | 70 |
126
- | RIOTING | 0.785 | 0.605 | 0.68 | 119 |
127
- | SEXUAL ASSAULT - OTHER | 0.829 | 0.839 | 0.834 | 971 |
128
- | SIMPLE ASSAULT | 0.977 | 0.967 | 0.972 | 4577 |
129
- | STOLEN PROPERTY - RECEIVING | 0.953 | 0.955 | 0.954 | 1193 |
130
- | STOLEN PROPERTY - TRAFFICKING | 0.899 | 0.875 | 0.887 | 491 |
131
- | TAX LAW (FEDERAL ONLY) | 0.474 | 0.177 | 0.256 | 30 |
132
- | TRAFFIC OFFENSES - MINOR | 0.976 | 0.975 | 0.975 | 8699 |
133
- | TRAFFICKING - COCAINE OR CRACK | 0.893 | 0.944 | 0.918 | 185 |
134
- | TRAFFICKING - DRUG UNSPECIFIED | 0.729 | 0.783 | 0.755 | 516 |
135
- | TRAFFICKING - HEROIN | 0.874 | 0.902 | 0.887 | 54 |
136
- | TRAFFICKING - OTHER CONTROLLED SUBSTANCES | 0.963 | 0.953 | 0.958 | 2832 |
137
- | TRAFFICKING MARIJUANA/HASHISH | 0.919 | 0.934 | 0.926 | 255 |
138
- | TRESPASSING | 0.974 | 0.982 | 0.978 | 1916 |
139
- | UNARMED ROBBERY | 0.941 | 0.935 | 0.938 | 377 |
140
- | UNAUTHORIZED USE OF VEHICLE | 0.929 | 0.911 | 0.92 | 304 |
141
- | UNSPECIFIED HOMICIDE | 0.641 | 0.591 | 0.614 | 60 |
142
- | VIOLENT OFFENSES - OTHER | 0.82 | 0.817 | 0.818 | 606 |
143
- | VOLUNTARY/NONNEGLIGENT MANSLAUGHTER | 0.641 | 0.559 | 0.596 | 54 |
144
- | WEAPON OFFENSE | 0.944 | 0.947 | 0.945 | 2466 |
145
 
146
  *Note: `support` is the average number of observations predicted on per fold, so the total number of observations per class is roughly 3x `support`.*
147
 
@@ -151,9 +151,9 @@ If we interpret the classification probability as a confidence score, we can use
151
 
152
  | | cutoff | percent retained | mcc | acc |
153
  | --- | ------ | ---------------- | ----- | ----- |
154
- | 0 | 0.85 | 0.952 | 0.959 | 0.961 |
155
- | 1 | 0.90 | 0.944 | 0.963 | 0.965 |
156
- | 2 | 0.95 | 0.928 | 0.969 | 0.971 |
157
  | 3 | 0.975 | 0.912 | 0.975 | 0.976 |
158
- | 4 | 0.99 | 0.885 | 0.982 | 0.983 |
159
- | 5 | 0.999 | 0.737 | 0.996 | 0.996 |
 
49
 
50
  | Metric | precision | recall | f1-score |
51
  | --------- | --------- | ------ | -------- |
52
+ | macro avg | 0.811 | 0.786 | 0.794 |
53
 
54
 
55
  *Note*: These are the average of the values *per fold*, so *macro avg* is the average of the macro average of all categories per fold.
 
58
 
59
  | Category | precision | recall | f1-score | support |
60
  | ------------------------------------------------------ | --------- | ------ | -------- | ------- |
61
+ | AGGRAVATED ASSAULT | 0.954 | 0.954 | 0.954 | 4085 |
62
+ | ARMED ROBBERY | 0.961 | 0.955 | 0.958 | 1021 |
63
+ | ARSON | 0.946 | 0.954 | 0.95 | 344 |
64
+ | ASSAULTING PUBLIC OFFICER | 0.914 | 0.905 | 0.909 | 588 |
65
+ | AUTO THEFT | 0.962 | 0.962 | 0.962 | 1660 |
66
+ | BLACKMAIL/EXTORTION/INTIMIDATION | 0.872 | 0.871 | 0.872 | 627 |
67
+ | BRIBERY AND CONFLICT OF INTEREST | 0.784 | 0.796 | 0.79 | 216 |
68
+ | BURGLARY | 0.979 | 0.981 | 0.98 | 2214 |
69
+ | CHILD ABUSE | 0.805 | 0.78 | 0.792 | 139 |
70
+ | COCAINE OR CRACK VIOLATION OFFENSE UNSPECIFIED | 0.827 | 0.815 | 0.821 | 47 |
71
+ | COMMERCIALIZED VICE | 0.818 | 0.788 | 0.802 | 666 |
72
+ | CONTEMPT OF COURT | 0.982 | 0.987 | 0.984 | 2952 |
73
+ | CONTRIBUTING TO DELINQUENCY OF A MINOR | 0.544 | 0.333 | 0.392 | 50 |
74
+ | CONTROLLED SUBSTANCE - OFFENSE UNSPECIFIED | 0.864 | 0.791 | 0.826 | 280 |
75
  | COUNTERFEITING (FEDERAL ONLY) | 0 | 0 | 0 | 2 |
76
+ | DESTRUCTION OF PROPERTY | 0.97 | 0.968 | 0.969 | 2560 |
77
+ | DRIVING UNDER INFLUENCE - DRUGS | 0.567 | 0.603 | 0.581 | 34 |
78
+ | DRIVING UNDER THE INFLUENCE | 0.951 | 0.946 | 0.949 | 2195 |
79
+ | DRIVING WHILE INTOXICATED | 0.986 | 0.981 | 0.984 | 2391 |
80
+ | DRUG OFFENSES - VIOLATION/DRUG UNSPECIFIED | 0.903 | 0.911 | 0.907 | 3100 |
81
+ | DRUNKENNESS/VAGRANCY/DISORDERLY CONDUCT | 0.856 | 0.861 | 0.858 | 380 |
82
+ | EMBEZZLEMENT | 0.865 | 0.759 | 0.809 | 100 |
83
  | EMBEZZLEMENT (FEDERAL ONLY) | 0 | 0 | 0 | 1 |
84
+ | ESCAPE FROM CUSTODY | 0.988 | 0.991 | 0.989 | 4035 |
85
+ | FAMILY RELATED OFFENSES | 0.739 | 0.773 | 0.755 | 442 |
86
+ | FELONY - UNSPECIFIED | 0.692 | 0.735 | 0.712 | 122 |
87
+ | FLIGHT TO AVOID PROSECUTION | 0.46 | 0.407 | 0.425 | 38 |
88
+ | FORCIBLE SODOMY | 0.82 | 0.8 | 0.809 | 76 |
89
  | FORGERY (FEDERAL ONLY) | 0 | 0 | 0 | 2 |
90
+ | FORGERY/FRAUD | 0.911 | 0.928 | 0.919 | 4687 |
91
  | FRAUD (FEDERAL ONLY) | 0 | 0 | 0 | 2 |
92
+ | GRAND LARCENY - THEFT OVER $200 | 0.957 | 0.973 | 0.965 | 2412 |
93
+ | HABITUAL OFFENDER | 0.742 | 0.627 | 0.679 | 53 |
94
+ | HEROIN VIOLATION - OFFENSE UNSPECIFIED | 0.879 | 0.811 | 0.843 | 24 |
95
+ | HIT AND RUN DRIVING | 0.922 | 0.94 | 0.931 | 303 |
96
+ | HIT/RUN DRIVING - PROPERTY DAMAGE | 0.929 | 0.918 | 0.923 | 362 |
97
+ | IMMIGRATION VIOLATIONS | 0.84 | 0.609 | 0.697 | 19 |
98
+ | INVASION OF PRIVACY | 0.927 | 0.923 | 0.925 | 1235 |
99
+ | JUVENILE OFFENSES | 0.928 | 0.866 | 0.895 | 144 |
100
+ | KIDNAPPING | 0.937 | 0.93 | 0.933 | 553 |
101
+ | LARCENY/THEFT - VALUE UNKNOWN | 0.955 | 0.945 | 0.95 | 3175 |
102
+ | LEWD ACT WITH CHILDREN | 0.775 | 0.85 | 0.811 | 596 |
103
+ | LIQUOR LAW VIOLATIONS | 0.741 | 0.768 | 0.755 | 214 |
104
+ | MANSLAUGHTER - NON-VEHICULAR | 0.626 | 0.802 | 0.701 | 139 |
105
+ | MANSLAUGHTER - VEHICULAR | 0.79 | 0.853 | 0.819 | 117 |
106
+ | MARIJUANA/HASHISH VIOLATION - OFFENSE UNSPECIFIED | 0.741 | 0.662 | 0.699 | 62 |
107
+ | MISDEMEANOR UNSPECIFIED | 0.63 | 0.243 | 0.347 | 57 |
108
+ | MORALS/DECENCY - OFFENSE | 0.774 | 0.764 | 0.769 | 412 |
109
+ | MURDER | 0.965 | 0.915 | 0.939 | 621 |
110
+ | OBSTRUCTION - LAW ENFORCEMENT | 0.939 | 0.947 | 0.943 | 4220 |
111
+ | OFFENSES AGAINST COURTS, LEGISLATURES, AND COMMISSIONS | 0.881 | 0.895 | 0.888 | 1965 |
112
+ | PAROLE VIOLATION | 0.97 | 0.953 | 0.962 | 946 |
113
+ | PETTY LARCENY - THEFT UNDER $200 | 0.965 | 0.761 | 0.85 | 139 |
114
+ | POSSESSION/USE - COCAINE OR CRACK | 0.893 | 0.928 | 0.908 | 68 |
115
+ | POSSESSION/USE - DRUG UNSPECIFIED | 0.624 | 0.535 | 0.572 | 189 |
116
+ | POSSESSION/USE - HEROIN | 0.884 | 0.852 | 0.866 | 25 |
117
+ | POSSESSION/USE - MARIJUANA/HASHISH | 0.977 | 0.97 | 0.973 | 556 |
118
+ | POSSESSION/USE - OTHER CONTROLLED SUBSTANCES | 0.975 | 0.965 | 0.97 | 3271 |
119
+ | PROBATION VIOLATION | 0.963 | 0.953 | 0.958 | 1158 |
120
+ | PROPERTY OFFENSES - OTHER | 0.901 | 0.87 | 0.885 | 446 |
121
+ | PUBLIC ORDER OFFENSES - OTHER | 0.7 | 0.721 | 0.71 | 1871 |
122
  | RACKETEERING/EXTORTION (FEDERAL ONLY) | 0 | 0 | 0 | 2 |
123
+ | RAPE - FORCE | 0.842 | 0.873 | 0.857 | 641 |
124
+ | RAPE - STATUTORY - NO FORCE | 0.707 | 0.55 | 0.611 | 140 |
125
+ | REGULATORY OFFENSES (FEDERAL ONLY) | 0.847 | 0.567 | 0.674 | 70 |
126
+ | RIOTING | 0.784 | 0.605 | 0.68 | 119 |
127
+ | SEXUAL ASSAULT - OTHER | 0.836 | 0.836 | 0.836 | 971 |
128
+ | SIMPLE ASSAULT | 0.976 | 0.967 | 0.972 | 4577 |
129
+ | STOLEN PROPERTY - RECEIVING | 0.959 | 0.957 | 0.958 | 1193 |
130
+ | STOLEN PROPERTY - TRAFFICKING | 0.902 | 0.888 | 0.895 | 491 |
131
+ | TAX LAW (FEDERAL ONLY) | 0.373 | 0.233 | 0.286 | 30 |
132
+ | TRAFFIC OFFENSES - MINOR | 0.974 | 0.977 | 0.976 | 8699 |
133
+ | TRAFFICKING - COCAINE OR CRACK | 0.896 | 0.951 | 0.922 | 185 |
134
+ | TRAFFICKING - DRUG UNSPECIFIED | 0.709 | 0.795 | 0.749 | 516 |
135
+ | TRAFFICKING - HEROIN | 0.871 | 0.92 | 0.894 | 54 |
136
+ | TRAFFICKING - OTHER CONTROLLED SUBSTANCES | 0.963 | 0.954 | 0.959 | 2832 |
137
+ | TRAFFICKING MARIJUANA/HASHISH | 0.921 | 0.943 | 0.932 | 255 |
138
+ | TRESPASSING | 0.974 | 0.98 | 0.977 | 1916 |
139
+ | UNARMED ROBBERY | 0.941 | 0.939 | 0.94 | 377 |
140
+ | UNAUTHORIZED USE OF VEHICLE | 0.94 | 0.908 | 0.924 | 304 |
141
+ | UNSPECIFIED HOMICIDE | 0.61 | 0.554 | 0.577 | 60 |
142
+ | VIOLENT OFFENSES - OTHER | 0.827 | 0.817 | 0.822 | 606 |
143
+ | VOLUNTARY/NONNEGLIGENT MANSLAUGHTER | 0.619 | 0.513 | 0.542 | 54 |
144
+ | WEAPON OFFENSE | 0.943 | 0.949 | 0.946 | 2466 |
145
 
146
  *Note: `support` is the average number of observations predicted on per fold, so the total number of observations per class is roughly 3x `support`.*
147
 
 
151
 
152
  | | cutoff | percent retained | mcc | acc |
153
  | --- | ------ | ---------------- | ----- | ----- |
154
+ | 0 | 0.85 | 0.952 | 0.96 | 0.961 |
155
+ | 1 | 0.9 | 0.943 | 0.964 | 0.965 |
156
+ | 2 | 0.95 | 0.928 | 0.97 | 0.971 |
157
  | 3 | 0.975 | 0.912 | 0.975 | 0.976 |
158
+ | 4 | 0.99 | 0.886 | 0.982 | 0.983 |
159
+ | 5 | 0.999 | 0.733 | 0.995 | 0.996 |
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c863f2c55a71623041440ab4366ed794ef4f6d874805adf05bee940c126b2b10
3
  size 331143181
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:89d279378f136333c365d6d0d13a79e7b0915b34c296b26755fdc4a84fe19086
3
  size 331143181