# Copyright Redis Ltd. 2021 - present # Licensed under your choice of the Redis Source Available License 2.0 (RSALv2) or # the Server Side Public License v1 (SSPLv1). import matplotlib.pyplot as plt import numpy as np from sklearn.tree import DecisionTreeClassifier from sklearn import tree '''------------------- batches HNSW -----------------''' def print_tree(clf): tree.plot_tree(clf) plt.savefig('batches_vs_adhoc-HNSW.pdf') plt.show() n_nodes = clf.tree_.node_count children_left = clf.tree_.children_left children_right = clf.tree_.children_right feature = clf.tree_.feature threshold = clf.tree_.threshold node_depth = np.zeros(shape=n_nodes, dtype=np.int64) is_leaves = np.zeros(shape=n_nodes, dtype=bool) stack = [(0, 0)] # start with the root node id (0) and its depth (0) while len(stack) > 0: # `pop` ensures each node is only visited once node_id, depth = stack.pop() node_depth[node_id] = depth # If the left and right child of a node is not the same we have a split # node is_split_node = children_left[node_id] != children_right[node_id] # If a split node, append left and right children and depth to `stack` # so we can loop through them if is_split_node: stack.append((children_left[node_id], depth + 1)) stack.append((children_right[node_id], depth + 1)) else: is_leaves[node_id] = True print("The binary tree structure has {n} nodes and has the following tree structure:\n".format(n=n_nodes)) for i in range(n_nodes): if is_leaves[i]: print("{space}node={node} is a leaf node.".format(space=node_depth[i] * "\t", node=i)) else: print("{space}node={node} is a split node: go to node {left} if X[:, {feature}] <= {threshold} " "else to node {right}.".format(space=node_depth[i] * "\t", node=i, left=children_left[i], feature=feature[i], threshold=threshold[i], right=children_right[i])) if __name__ == '__main__': clf = DecisionTreeClassifier(max_leaf_nodes=20) # features: (k, index_size, dim, M, r) # labels: 1: adhoc better, -1: batches better y = [] # The following labels represent the results for experiments that were held for the following combinations: # k = 5, 50, 500 # index_sizes: 1000, 10000, 100000, 1M, 10M # dim: 10, 100, 1000 (except for size=1M or 10M with dim==100,1000) # M: 4, 16, 64 # ratio: 0.02-0.2 with steps of 0.02 X_1 = np.array([[5*k, 10**i, 10**j, 4**m, 2*l/100] for k in range(1, 4) for i in range(3, 8) for j in range(1, 4) for m in range(1, 4) for l in range(1, 21) if i+j <= 8 and not (i == 6 and j == 2)]) # (5, 1000, 10, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # (5, 1000, 100, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # (5, 1000, 1000, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # (5, 10000, 10, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) # (5, 10000, 100, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, -1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1]) # (5, 10000, 1000, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # (5, 100k, 10, [4,16,64], [0.02-0.2]) y.extend([1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) # (5, 100k, 100, [4,16,64], [0.02-0.2]) y.extend([1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) # (5, 100k, 1000, [4,16,64], [0.02-0.2]) y.extend([1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) # (5, 1M, 10, [4,16,64], [0.02-0.2]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) # (5, 10M, 10, [4,16,64], [0.02-0.2]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) # (50, 1000, 10, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # (50, 1000, 100, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # (50, 1000, 1000, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # (50, 10000, 10, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) # (50, 10000, 100, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # (50, 10000, 1000, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # (50, 100k, 10, [4,16,64], [0.02-0.2]) y.extend([1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) # (50, 100k, 100, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) # (50, 100k, 1000, [4,16,64], [0.02-0.2]) y.extend([1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1]) # (50, 1M, 10, [4,16,64], [0.02-0.2]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) # (50, 10M, 10, [4,16,64], [0.02-0.2]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) # (500, 1000, 10, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # (500, 1000, 100, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # (500, 1000, 1000, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # (500, 10000, 10, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1]) # (500, 10000, 100, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # (500, 10000, 1000, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # (500, 100k, 10, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) # (500, 100k, 100, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # (500, 100k, 1000, [4,16,64], [0.02-0.2]) y.extend([1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # (500, 1M, 10, [4,16,64], [0.02-0.2]) y.extend([1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) # (500, 10M, 10, [4,16,64], [0.02-0.2]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) # The following labels represent the results for experiments that were held for the following combinations: # k = 5, 50, 500 # index_sizes: 1M # dim: 100 # M: 50 # ratio: 0.02-0.2 with steps of 0.02 X_2 = np.array([[5*k, 1e6, 100, 50, 2*l/100] for k in range(1, 4) for l in range(1, 11)]) # (5, 1M, 100, 50, [0.02-0.2]) y.extend([1, 1, -1, -1, -1, -1, -1, -1, -1, -1]) # (50, 1M, 100, 50, [0.02-0.2]) y.extend([1, 1, 1, -1, -1, -1, -1, -1, -1, -1]) # (500, 1M, 100, 50, [0.02-0.2]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, -1, -1]) # The following labels represent the results for experiments that were held for the following combinations: # k = 5, 50, 500 # index_sizes: 50000, 500000, 5M # dim: 5, 50, 500 (except for size=5M with dim=500) # M: 24, 48 # ratio: 0.05-0.5 with steps of 0.05 X_3 = np.array([[5*k, 5*10**i, 5*10**j, m, 5*l/100] for i in range(4, 7) for j in range(3) for m in [24, 48] for k in range(1, 4) for l in range(1, 11) if i+j < 8]) # ([5,50,500], 50k, 5, 24, [0.05-0.5]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, -1, -1, -1, -1, -1, -1, -1]) # ([5,50,500], 50k, 5, 48, [0.05-0.5]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, -1, -1, -1, -1, -1, -1, -1]) # ([5,50,500], 50k, 50, 24, [0.05-0.5]) y.extend([1, 1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, -1, -1]) # ([5,50,500], 50k, 50, 48, [0.05-0.5]) y.extend([1, 1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, -1]) # ([5,50,500], 50k, 500, 24, [0.05-0.5]) y.extend([1, 1, 1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # ([5,50,500], 50k, 50, 48, [0.05-0.5]) y.extend([1, 1, 1, 1, 1, 1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, -1]) # ([5,50,500], 0.5M, 5, 24, [0.05-0.5]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) # ([5,50,500], 0.5M, 5, 48, [0.05-0.5]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) # ([5,50,500], 0.5M, 50, 24, [0.05-0.5]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, -1, -1, -1, -1, -1, -1, -1]) # ([5,50,500], 0.5M, 50, 48, [0.05-0.5]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, -1, -1, -1, -1, -1, -1, -1]) # ([5,50,500], 0.5M, 500, 24, [0.05-0.5]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, -1, -1, -1, -1, -1, -1, -1]) # ([5,50,500], 0.5M, 500, 48, [0.05-0.5]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, -1, -1, -1, -1, -1, -1]) # ([5,50,500], 5M, 5, 24, [0.05-0.5]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) # ([5,50,500], 5M, 5, 48, [0.05-0.5]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) # ([5,50,500], 5M, 50, 24, [0.05-0.5]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) # ([5,50,500], 5M, 50, 48, [0.05-0.5]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([-1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) # The following labels represent the results for experiments that were held for the following combinations: # k = 5, 50, 500 # index_sizes: 500000, 5M # dim: 5, 50, 500 (except for size=5M with dim=500) # M: 24, 48 # ratio: 0.002-0.02 with steps of 0.002 X_4 = np.array([[5*k, 5*10**i, 5*10**j, m, 2*l/1000] for i in range(5, 7) for j in range(3) for m in [24, 48] for k in range(1, 4) for l in range(1, 11) if i+j < 8]) # ([5,50,500], 0.5M, 5, 24, [0.002-0.02]) y.extend([1, 1, 1, 1, 1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # ([5,50,500], 0.5M, 5, 48, [0.002-0.02]) y.extend([1, 1, 1, 1, 1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # ([5,50,500], 0.5M, 50, 24,[0.002-0.02]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # ([5,50,500], 0.5M, 50, 48, [0.002-0.02]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # ([5,50,500], 0.5M, 500, 24, [0.002-0.02]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # ([5,50,500], 0.5M, 500, 48, [0.002-0.02]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # ([5,50,500], 5M, 5, 24, [0.002-0.02]) y.extend([1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # ([5,50,500], 5M, 5, 48, [0.002-0.02]) y.extend([1, -1, -1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # ([5,50,500], 5M, 50, 24, [0.002-0.02]) y.extend([1, 1, 1, -1, -1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # ([5,50,500], 5M, 50, 48, [0.002-0.02]) y.extend([1, 1, 1, 1, 1, -1, -1, -1, -1, -1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) y.extend([1, 1, 1, 1, 1, 1, 1, 1, 1, 1]) # The total train set is the concatenation of the 4 combinations sets above. X = np.concatenate((X_1, X_2, X_3, X_4)) clf = clf.fit(X, y) print_tree(clf) # (k, index_size, dim, M, r) print(clf.predict([[10, 800000, 100, 50, 0.0001]]))