DexGroves

# Make a 3D transition matrix T where z-dimension is time
T_states <- 3
T_terms <- 4
T <- array(runif(T_states * T_states * Tz)/1.5, dim = c(T_states, T_states, Tz))

# Normalise by rows 
for (z in seq(Tz)) {
  T[, , z] <- T[, , z] / apply(T[, , z], 1, sum)
}

DexGroves

for (i in 2:ncol(my_data)) {
  varname <- colnames(my_data)[i]
  # if you want to change varname somehow
  # varname <- paste0(varname, "new")
  my_data[[varname]] <- my_data[,i]/sd(my_data[,i])
}

DexGroves

library("randomForest")
data(mtcars)

rf <- randomForest(mpg ~ ., data = mtcars, ntree = 10)
preds <- predict(rf, newdata = mtcars, predict.all = TRUE)

# Aggregate predictions 
preds$aggregate
# Invididual tree predictions
preds$individual 

DexGroves

M1 <- matrix(seq(16), nrow = 4, ncol = 4)
M2 <- matrix(seq(16), nrow = 4, ncol = 4)

M1 * M2 
M1 %*% M2 

`*` <- function(lhs, rhs) {
  if (is.matrix(lhs) & is.matrix(rhs)) {
    return(lhs %*% rhs)
  }

DexGroves

import numpy as np
from collections import deque


def dist_to(game_map, sources):
    dist = np.empty_like(game_map.owners)
    dist.fill(-1)

    for sx, sy in sources:
        dist[sx, sy] = 0

DexGroves

import numpy as np
import pandas as pd
from sklearn.linear_model import LinearRegression

N = 10_000
np.random.seed(42)

df = pd.DataFrame(index=np.arange(N))

df['x2'] = np.random.choice([1,2,3], size=N)