Chapter 8, Putting Things Away

1. From the beginning to the end (start to goal); from goal to start; and from both ends at once and meeting in the middle.
2. By eliminating the effect of the item on the branch. For example, using our "robot does not move" fault, if the branch says, "Arduino- no power", if you check to see if the Arduino has power, and it does, you can prune that branch. If the branch is “Motor stuck”, the effect of having a motor stuck is that the robot will drive in circles. As the robot is not driving in circles – it’s not driving at all – you can prune that branch.
3. It determines the amount of “impurity” in the sample or pool. When the Gini impurity = 0, all of the members of the class have the same attributes, and no further subdivision is possible. Gini minimizes misclassification. The Gini Index is 1 minus the sum of the squares of the probability of an item being in that class.

4. Color, noise, soft, and material were not useful in dividing the categories by labels. The labels and the items did not correlate. It does make sense that color is not useful in subdividing toys by type.
5. The color white was used by the decision tree that used the Gini Index and One Hot Encoding to separate out the stuffed animals.
6. Let’s have three types of menu items: appetizers, entrée, and dessert. Label encoding would substitute 0 for appetizers, 1 for entrée, and 2 for dessert. One Hot Encoding would use 1 0 0 for appetizers, 0 1 0 for entrees, and 0 0 1 for desserts.
7. The G() function is the distance along the path from the current position to the start. H() is the distance from the current position directly to the goal (Euclidean Distance). Note that G() follows the path, and H() is the straight line distance to the goal, since we have not computed a path to the goal yet.
8. A heuristic is an approach that is not guaranteed to be optimal, but instead is just sufficient for the task. Since H(), the direct line distance to the goal, is an estimate and ignores any obstacles, it can’t be used directly, but is just a way to compare one position to another. A major difference between D* and A* is that D* starts at the goal and works backwards toward the start. This allows D* to know the exact cost to the target – it is using the actual path distance to the goal from the current position and not a heuristic or estimate of the distance to go, like A* did.
9. RAISED squares or points are eliminated from consideration. LOWERED squares may be added back into the queue for consideration to be a path. Keep in mind that lowering scores due to new sensor readings ripple through the path planner.