-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathThis_wan.py
186 lines (153 loc) · 6.18 KB
/
This_wan.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
import pandas as pd
import typing
import matplotlib.pyplot as plt
import numpy as np
from sklearn import model_selection
from sklearn import linear_model
from sklearn import preprocessing
event_data = pd.read_csv("C:/Users/Jack/Documents/Coding Projects/UFC/Data/ufc_event_data.csv")
fight_data = pd.read_csv("C:/Users/Jack/Documents/Coding Projects/UFC/Data/ufc_fight_data.csv")
fight_stat_data = pd.read_csv("C:/Users/Jack/Documents/Coding Projects/UFC/Data/ufc_fight_stat_data.csv")
fighter_data = pd.read_csv("C:/Users/Jack/Documents/Coding Projects/UFC/Data/ufc_fighter_data.csv")
def stat_prcnt_calc(df: pd.DataFrame, att_var: str, succ_var: str) -> pd.Series:
return df[succ_var]/df[att_var]
def row_sum_ignore_na(row: pd.Series, cols: list) -> float:
if all([pd.isna(row[i]) for i in cols]): # If all values are na return na
return np.nan
res = 0
for i in cols:
if pd.isna(row[i]) == False:
res += row[i]
return res
def min_sec_to_sec(row: pd.Series, time: str) -> float:
try:
mins_secs = row[time].split(":")
except:
return np.nan
else:
if len(mins_secs) == 2:
mins = int(mins_secs[0]) * 60
secs = int(mins_secs[1])
res = mins + secs
return res
else:
return np.nan
df = fight_stat_data.merge(fighter_data,
how = "left",
left_on = "fighter_id",
right_on = "fighter_id")
df["strike_succ_prcnt"] = stat_prcnt_calc(df,"total_strikes_att", "total_strikes_succ")
df["sig_strike_succ_prcnt"] = stat_prcnt_calc(df,"sig_strikes_att", "sig_strikes_succ")
df["takedown_succ_prcnt"] = stat_prcnt_calc(df,"takedown_att", "takedown_succ")
# Calculating total number of fights - using row_sum_ignore_na
df["tot_fights"] = df.apply(
lambda row: row_sum_ignore_na(
row, ["fighter_w", "fighter_l", "fighter_d", "fighter_nc_dq"]),
axis = 1)
# Calculating total ufc fights by counting rows when grouping by fighter_id
tot_ufc_fights = df.groupby("fighter_id").size().reset_index(name="Count")
# Joining with fighter_data to get name for fighter
tot_ufc_fights = tot_ufc_fights.merge(fighter_data[["fighter_id","fighter_f_name", "fighter_l_name"]], how = "left", left_on = "fighter_id", right_on = "fighter_id")
# Raw winner variable is as the fighter_id, want to conver this to a binary when joint on fighter_data
wins = fight_data[["fight_id",
"event_id",
"winner",
"result",
"result_details",
"finish_round",
"finish_time"]]
# Creating dependent variable by merging the above dataframe
df = df.merge(wins, how = "left", left_on = "fight_id", right_on = "fight_id")
df["won"] = df.apply(lambda row: 1 if row["fighter_id"] == row["winner"] else 0, axis = 1)
df.drop("winner", inplace = True, axis = 1) # We have replaced this with our binary "won"
# Converting ctrl_time and finish_time into a integer representation as seconds
df["ctrl_time_sec"] = df.apply(
lambda row: min_sec_to_sec(row, "ctrl_time"), axis = 1
)
df["finish_time_sec"] = df.apply(
lambda row: min_sec_to_sec(row, "finish_time"), axis = 1
)
df["result"] = df["result"].astype("category")
enc = preprocessing.OneHotEncoder(sparse_output=False)
enc_data = enc.fit_transform(df[["result"]])
enc_feature_names = enc.get_feature_names_out(["result"])
df1 = pd.concat([df, pd.DataFrame(enc_data, columns = enc_feature_names)], axis = 1)
df1_a = df1[df1["fight_id"].duplicated(keep="first")]
df1_b = df1[df1["fight_id"].duplicated(keep="last")]
vs_df = pd.merge(df1_a, df1_a, how="left", on="fight_id", suffixes=["_A", "_B"])
xs = df1.drop(["fighter_url",
"fight_url",
"fighter_nickname",
"event_id",
"fight_stat_id",
"fight_id",
"fighter_id",
"won",
"fighter_f_name",
"fighter_l_name",
"result",
"result_details",
"finish_time",
"ctrl_time"],
axis = 1)
xs = xs.apply(pd.to_numeric, errors = "coerce")
xs = (xs-xs.mean())/xs.std()
xs = xs.fillna(0)
y = df["won"]
xs = xs.fillna(0)
vaz = ["fighter_url",
"fight_url",
"fighter_nickname",
"event_id",
"fight_stat_id",
"fight_id",
"fighter_id",
"won",
"fighter_f_name",
"fighter_l_name",
"result",
"result_details",
"finish_time",
"ctrl_time"]
a_vars = [i + "_A" for i in vaz if i != "fight_id"]
b_vars = [i + "_B" for i in vaz if i != "fight_id" and "won_B"]
ab_vars = a_vars + b_vars + ["fight_id"]
xs = vs_df.drop(ab_vars, axis = 1)
y = vs_df["won_A"]
xs = xs.apply(pd.to_numeric, errors = "coerce")
xs = (xs-xs.mean())/xs.std()
xs = xs.fillna(0)
X_train, X_test, y_train, y_test = model_selection.train_test_split(
xs, y, test_size = 0.33, random_state=5)
# Modelling
model = linear_model.LogisticRegression(random_state=1,
max_iter = 500,
penalty = "elasticnet",
solver = "saga",
l1_ratio = 0.5
).fit(X_train, y_train)
preds = model.predict(X_test)
model.score(X_test, y_test)
# Think Bayesian makes more sense
# Fighter dict:
numeric_cols = pd.Series([i for i in df1.columns if df1[i].dtype in ["float64","int64"]])
fighters = df1[numeric_cols]
cols_to_avg = ["total_strikes_att",
"total_strikes_succ",
"sig_strikes_att",
"sig_strikes_succ",
"takedown_att",
"takedown_succ",
"submission_att",
"reversals",
"strike_succ_prcnt",
"sig_strike_succ_prcnt",
"takedown_succ_prcnt",
"ctrl_time_sec",
"finish_time_sec",
"result_DQ",
"result_Decision",
"result_KO/TKO",
"result_Submission",
"result_TKO - Doctor's Stoppage"]
fighters = fighters.groupby("fighter_id")[cols_to_avg].mean()