Unit 6-8
SafeAssign Originality Report
LDR 5301-22.01.00-1B23-S1, Methods of Analysis for Business Operations•Unit VI Case Study
Connie StanleyTotal Score:highRisk Created with Sketch.
Total Number of Reports
1Highest Match
42 % Unit 6 Sept 2022.docxAverage Match
42 %Submitted on
09/27/22 09:05 PM EDTAverage Word Count
1,267 Highest: Unit 6 Sept 2022.docxhighRisk Created with Sketch.
- Internet (4)
28%
- 2
educba
- 6
sbmu
- 1
worldwidescience
- 4
dokumen
- 2
- Global database (8)
14%
- 7
Student paper
- 3
Student paper
- 11
Student paper
- 9
Student paper
- 12
Student paper
- 5
Student paper
- 8
Student paper
- 10
Student paper
- 7
- Top sources (3)
- 2
educba
- 7
Student paper
- 3
Student paper
- 2
- Excluded sources (0)
Guy Josobo’s Dilemma
Columbia Southern University
Connie Stanley
September 27, 2022
1The decision tree
Figure 1:
The decision tree to solve Guy Josobo’s dilemma
The best decision alternative
The best decision alternative for making their battery and outsourcing is going with the medium demand (Barrett, Lentz, & Maxwell, 2017).
For both alternatives, the medium demand does not lead to losses although the profits are not as high.
State probability
The state probability is as shown in the figure above
Cost at each income node
The cost matrix is as follows.
Now calculating the costs
Costs = * =
From the above calculations, the option that brings the best outcome is making their own when the demand is highest in the market (Bhukya & Ramachandram, 2017).
The revenue generated than is 25000.
The other best alternative is outsourcing when the demand is medium in the market.
Challenges faced when using the decision tree
Regression is not recommended
Using the logarithmic regression approach, we can predict with precision the probability of outcome using features independent of each other (C, & Babu, 2016).
The model may malfunction when a large amount of data is used for training, which can cause it to overestimate the prediction accuracy on the training dataset, hindering the model from accurately predicting the results on the test dataset when high-dimensional datasets are used (Weinberg, & Last, 2019).
Models with many features are most likely to suffer from this if they have been trained on a small amount of data (Weinberg, & Last, 2019).
2To minimize over-fitting on high-dimensional datasets, regularization strategies should be considered (but this makes the model more complex). In the case of too high regularization variables, the model may not fit the training data well (Weinberg, & Last, 2019).
The logistic regression method has difficulty capturing complex correlations.
A more powerful algorithm, such as a neural network, can easily outperform this approach (Bhukya & Ramachandram, 2017).
Logic regression does not handle nonlinear issues because of its linear decision surface.
Data that can be linearly separated is rare in real-world circumstances (Bhukya & Ramachandram, 2017).
It is then necessary to transform non-linear features into linear features.
2This can be done by increasing the number of features to allow the data to be divided linearly into higher dimensions as a result of the non-linear features (Weinberg, & Last, 2019).
The problem of overfitting
2The decision-tree learner is unable to generalize the input well, so overly complicated trees can be created (Bhukya & Ramachandram, 2017).
Overfitting is a term used to describe this.
This can be avoided by using the following techniques:
Performing pruning
When the tree is developed further, each row of the input data table may be the last rule of the tree (C, & Babu, 2016).
2When the model is designed for training data, it will perform perfectly, but when the model is created for test data, it will fail to validate.
It is important to note that overfitting occurs as soon as the complexity of the algorithm reaches a certain level (Bhukya & Ramachandram, 2017).
The likelihood of overfitting is quite high in large trees that have many nodes.
Attempts are made to avoid overfitting in the decision.
There are nearly always going to be no leaf nodes before the tree reaches depth;
2thus, every leaf node will only have observations from one observation point or one class.
It is possible to determine the correct time for stopping the growth of the tree in several different ways (Bhukya & Ramachandram, 2017).
2It is very likely that trees downstream from a leaf node will not grow from that node if it is a pure node at any stage of its growth process (Bhukya & Ramachandram, 2017). To continue growing the tree, the tree can use other leaf nodes to continue growing.
Impurity levels decrease relatively slowly when trees become less impurity prone.
2This user input parameter terminates the tree if the impurity drops by less than 0.001, say (Weinberg, & Last, 2019). The leaf node can be considered complete when only a few observations are left on the leaf node.
As a result, the tree will be terminated when the reliability of a node for further splitting is questioned due to the small sample size because it is only small sample size (Bhukya & Ramachandram, 2017).
2When they are mutually independent, a large sample is composed of around thirty observations based on the Central Limit Theorem.
To provide a general guide, let us say that this is a reasonable level of user input parameter for multidimensional data (Weinberg, & Last, 2019).
However, because we work with multidimensional data that may be associated, the level should be higher than 30, for example, 50, 100, or 200 (Barrett, Lentz, & Maxwell, 2017).
Expensive
2Because each node in a decision tree requires field sorting, the cost of creating one is high.
Another algorithm involves using several fields at once, increasing costs even further (TOFAN, 2014).
It is also expensive to use pruning methods since there are many candidate subtrees to produce and compare.
Analyses that are independent of one another
There must be total independence between each training example in a dataset (TOFAN, 2014).
2Those specific training instances will be given more weight if they are related in some way. To avoid duplicated measurements or matched data, training data should not be matched.
There is instability
Instabilities in decision trees are caused by subtle changes in the data that can change the tree's structure entirely (Weinberg, & Last, 2019).
By incorporating decision trees into an ensemble, this difficulty can be alleviated (Weinberg, & Last, 2019).
Greedy behavior
2Creating a binary tree requires a correctly partitioned input space.
For this, greedy algorithms are used to recursively split binary files.
Various values must be aligned in this numerical procedure (Weinberg, & Last, 2019).
There will only be one path that will be taken to split the data.
2The split of the data will be determined according to the first best split.
Despite this, there may be more effective ways of splitting the data, so they will not be the best path to take;
when reading the data (Bhukya & Ramachandram, 2017).
What decision should be made
The best decision the Guy should make is to make their battery.
While the risks are quite high, the rewards are equally high when the demand in the market is high (Bhukya & Ramachandram, 2017).
Besides the alternative is to outsource where the only best outcoming is not making any loss but the income is comparatively low in the long run.
References
Barrett, C.
3B., Lentz, E., & Maxwell, D.
G.
(2017).
3A Market Analysis and Decision Tree Tool for Response Analysis: Cash, Local Purchase, and/or Imported Food Aid? The Decision Tree Tool. SSRN Electronic Journal.
4https://doi.org/10.2139/ssrn.1141992
Bhukya, D.
5P., & Ramachandram, S.
(2017).
6Decision Tree Induction: An Approach for Data Classification Using AVL-Tree.
7International Journal of Computer and Electrical Engineering, 660–665.
8https://doi.org/10.7763/ijcee.2010.v2.208
C, K.
K.
R., & Babu, V.
(2016).
A Survey on Issues of Decision Tree and Non-Decision Tree Algorithms.
International Journal of Artificial Intelligence and Applications for Smart Devices, 4(1), 9–32.
https://doi.org/10.14257/ijaiasd.2016.4.1.02
TOFAN, C.
A.
(2014).
7Optimization Techniques of Decision Making - Decision Tree. Advances in Social Sciences Research Journal, 1(5), 142–148.
9https://doi.org/10.14738/assrj.15.437
Weinberg, A.
10I., & Last, M.
(2019).
11Selecting a representative decision tree from an ensemble of decision-tree models for fast big data classification.
12Journal of Big Data, 6(1).
11https://doi.org/10.1186/s40537-019-0186-3
Source Matches (26)- 1worldwidescience100%
Student paper
The decision tree
Original source
The Decision Tree
- 2educba82%
Student paper
To minimize over-fitting on high-dimensional datasets, regularization strategies should be considered (but this makes the model more complex). In the case of too high regularization variables, the model may not fit the training data well (Weinberg, & Last, 2019).
Original source
Regularization strategies should be considered on high-dimensional datasets to minimize over-fitting (but this makes the model complex) The model may be under-fit on the training data if the regularization variables are too high
- 2educba72%
Student paper
This can be done by increasing the number of features to allow the data to be divided linearly into higher dimensions as a result of the non-linear features (Weinberg, & Last, 2019).
Original source
As a result, non-linear features must be transformed, which can be done by increasing the number of features such that the data can be separated linearly in higher dimensions
- 2educba68%
Student paper
The decision-tree learner is unable to generalize the input well, so overly complicated trees can be created (Bhukya & Ramachandram, 2017).
Original source
Overly complicated trees can be created by decision-tree learners, which do not generalize the input well
- 2educba70%
Student paper
When the model is designed for training data, it will perform perfectly, but when the model is created for test data, it will fail to validate.
Original source
On the training data, the model will perform admirably, but it will fail to validate on the test data
- 2educba84%
Student paper
thus, every leaf node will only have observations from one observation point or one class.
Original source
thus, each leaf node only includes observations from one class or one observation point
- 2educba72%
Student paper
It is very likely that trees downstream from a leaf node will not grow from that node if it is a pure node at any stage of its growth process (Bhukya & Ramachandram, 2017). To continue growing the tree, the tree can use other leaf nodes to continue growing.
Original source
If a leaf node is a pure node at any point during the growth process, no additional downstream trees will grow from that node Other leaf nodes can be used to continue growing the tree
- 2educba64%
Student paper
This user input parameter terminates the tree if the impurity drops by less than 0.001, say (Weinberg, & Last, 2019). The leaf node can be considered complete when only a few observations are left on the leaf node.
Original source
When the impurity lowers by a very little amount, say 0.001 or less, this user input parameter causes the tree to be terminated When there are only a few observations remaining on the leaf node
- 2educba72%
Student paper
When they are mutually independent, a large sample is composed of around thirty observations based on the Central Limit Theorem.
Original source
According to the Central Limit Theorem, a big sample consists of around 30 observations when they are mutually independent
- 2educba86%
Student paper
Because each node in a decision tree requires field sorting, the cost of creating one is high.
Original source
Expensive The cost of creating a decision tree is high since each node requires field sorting
- 2educba70%
Student paper
Those specific training instances will be given more weight if they are related in some way. To avoid duplicated measurements or matched data, training data should not be matched.
Original source
If they are related in some manner, the model will try to give those specific training instances more weight As a result, no matched data or repeated measurements should be used as training data
- 2educba68%
Student paper
Creating a binary tree requires a correctly partitioned input space.
Original source
Greedy Approach To form a binary tree, the input space must be partitioned correctly
- 2educba66%
Student paper
The split of the data will be determined according to the first best split.
Original source
Data will be split according to the first best split, and only that path will be used to split the data
- 3Student paper100%
Student paper
B., Lentz, E., & Maxwell, D.
Original source
B., Lentz, E., & Maxwell, D
- 3Student paper95%
Student paper
A Market Analysis and Decision Tree Tool for Response Analysis: Cash, Local Purchase, and/or Imported Food Aid? The Decision Tree Tool. SSRN Electronic Journal.
Original source
A Market Analysis and Decision Tree Tool for Response Analysis Cash, Local Purchase and/or Imported Food Aid The Decision Tree Tool SSRN Electronic Journal SSRN
- 4dokumen71%
Student paper
https://doi.org/10.2139/ssrn.1141992
Original source
https://doi.org/10
- 5Student paper100%
Student paper
P., & Ramachandram, S.
Original source
P., & Ramachandram, S
- 6sbmu100%
Student paper
Decision Tree Induction: An Approach for Data Classification Using AVL-Tree.
Original source
Decision tree induction an approach for data classification using AVL-tree
- 7Student paper100%
Student paper
International Journal of Computer and Electrical Engineering, 660–665.
Original source
International Journal of Computer and Electrical Engineering, 660-665
- 8Student paper79%
Student paper
https://doi.org/10.7763/ijcee.2010.v2.208
Original source
10.7763/ijcee.2010.v2.208
- 7Student paper100%
Student paper
Optimization Techniques of Decision Making - Decision Tree. Advances in Social Sciences Research Journal, 1(5), 142–148.
Original source
Optimization Techniques of Decision Making - Decision Tree Advances in Social Sciences Research Journal, 1(5), 142-148
- 9Student paper100%
Student paper
https://doi.org/10.14738/assrj.15.437
Original source
https://doi.org/10.14738/assrj.15.437
- 10Student paper100%
Student paper
I., & Last, M.
Original source
I., & Last, M
- 11Student paper100%
Student paper
Selecting a representative decision tree from an ensemble of decision-tree models for fast big data classification.
Original source
Selecting a representative decision tree from an ensemble of decision‑tree models for fast big data classification
- 12Student paper89%
Student paper
Journal of Big Data, 6(1).
Original source
Journal of Big Data, 6(1), 23
- 11Student paper72%
Student paper
https://doi.org/10.1186/s40537-019-0186-3
Original source
Weinberg and Last J Big Data (2019) 6:23 https://doi.org/10.1186/s40537‑019‑0186‑3
- 1worldwidescience100%