Replicating a decision tree in tikz

1. Well, it's too wide, so you need to make things narrower. For example by reducing the s sep.

2. The color of the nodes and the color of the arrows are not connected, if a node is red it's because you've added ,red!70 for that specific node. So you just need to remove a lot of those ,red!70s.

3. You need to set the width and height separately:

     inner sep=0,
     minimum width=1em,
     minimum height=0.5em,
   

No inner sep, so there is no padding, then set minimum width/height to suitable values. You may want to modify them.

I also set s sep in the sample and feature bagging, node, to space the subtrees a bit more, and I added two phantom nodes for additional space between the second and third. And I added the dots later, by placing a node midway between t2 and tn.

I'd probably set the mean in regression.. node directly below t2, but I'll leave that for you to decide.

The frame in this screenshot is made by the showframe package, and indicates the width of the text block.

documentclass[]{article}

usepackage{
  forest,
 % showframe
 }
usetikzlibrary{fit,positioning}

tikzset{
  font=largesffamilybfseries,
  red arrow/.style={
    midway,red,sloped,fill, minimum height=3cm, single arrow, single arrow head extend=.5cm, single arrow head indent=.25cm,xscale=0.3,yscale=0.15,
    allow upside down
  },
  black arrow/.style 2 args={-stealth, shorten >=#1, shorten <=#2},
  black arrow/.default={1mm}{1mm},
  tree box/.style={draw, rounded corners, inner sep=1em},
  node box/.style={white, draw=black, text=black, rectangle, rounded corners},
}

begin{document}
begin{center}
begin{forest}
  for tree={
     l sep=2em,
     s sep=2mm,
     anchor=center,
     inner sep=0,
     minimum width=1em,
     minimum height=0.5em,
     fill=blue!50,
     rectangle,
     where level=2{no edge}{}}
  [
  Training Data, node box
  [sample and feature bagging, node box, alias=bagging, above=4em,s sep=1.1cm
  [,alias=a1[[,alias=a2][]][,edge label={node[above=1ex,red arrow]{}}[[][]]
  [,edge label={node[above=1ex,red arrow]{}}[,red!70,edge label={node[below=1ex,red arrow]{}}][,alias=a3]]]]
  [,alias=b1[,edge label={node[below=1ex,red arrow]{}}[[,alias=b2][]][,red!70,edge label={node[above=1ex,red arrow]{}}]][[][[][,alias=b3]]]]
  [,phantom]
  [,phantom]
  [,alias=c1[[,alias=c2][]][,edge label={node[above=1ex,red arrow]{}}[,edge label={node[above=1ex,red arrow]{}}[,alias=c3][,red!70,edge label={node[above=1ex,red arrow]{}}]][,alias=c4]]]]
  ]
  node[tree box, fit=(a1)(a2)(a3)](t1){};
  node[tree box, fit=(b1)(b2)(b3)](t2){};
  node[tree box, fit=(c1)(c2)(c3)(c4)](tn){};
  begin{scope}[every node/.append style={below right=0.5em, inner sep=0pt, font=normalsizesffamilybfseries}]
  node at (t1.north west) {Tree 1};
  node at (t2.north west) {Tree 2};
  node at (tn.north west) {Tree $n$};
  end{scope}
  path (t1.south west)--(tn.south east) node[midway,below=4em, node box] (mean) {mean in regression or majority vote in classification};
  node[below=3em of mean, node box] (pred) {prediction};
  draw[black arrow={5mm}{4mm}] (bagging) -- (t1.north);
  draw[black arrow] (bagging) -- (t2.north);
  draw[black arrow={5mm}{4mm}] (bagging) -- (tn.north);
  draw[black arrow={5mm}{5mm}] (t1.south) -- (mean);
  draw[black arrow] (t2.south) -- (mean);
  draw[black arrow={5mm}{5mm}] (tn.south) -- (mean);
  draw[black arrow] (mean) -- (pred);
  path (t2) -- node {dots} (tn); % <-- new node
end{forest}
end{center}
end{document}