Analysis of Various Decision Tree Algorithms for Classification in Data Mining
Citations
107 citations
62 citations
Cites background or methods from "Analysis of Various Decision Tree A..."
...The proposed deep learning architecture for tumor gene expression had achieved median testing accuracy with 96.90% in the 80% training and 20% testing strategy with BPSO-DT, which means that the using BPSO-DT algorithm and more data the architecture had, the better accuracy the architecture could achieve....
[...]
...According to the update mechanism in BPSO, each particle updates each position based on the particle’s velocity, which acts as probability threshold as shown in the VOLUME 8, 2020 22875 probabilistic equation (4). sig ( V ki ) = 1 1+ e− k i (3) X ki = { 1, if rand < sig(V ki ) 0, if otherwise (4) where 1 means this feature is chosen as an important feature for the next regeneration and 0 means this feature is not chosen as an important feature to the next regeneration, and rand is a random number ∈ [0, 1]....
[...]
...The steps of BPSO-DT are presented in Algorithm 1, where the input is the tumor gene expression dataset consisting of RNA-seq that needs optimization....
[...]
...Deep learning training is the third phase, which relies on the deep N. Algorithm 1 Extraction the Important Features of RNASeq Input : Tumor gene expression dataset Output: Gbest position 1 Initialize the position and velocity of each particle randomly 2 while iteration condition is not satisfied do 3 Evaluate the fitness of the particle swarm by DT according to equation 5 4 for each particle i do 5 if the fitness of xi is greater than the fitness of the Pbest i then 6 Pbest i = x i 7 end 8 if the fitness of any particle of the swarm is greater than Gbest then 9 Gbest = particle’s position 10 end 11 for each dimension D = 1, . . . ,N do 12 Update particles velocity and particles position according to equation 1,3, and 4 respectively 13 end 14 end 15 go to next generation until termination criterion is met 16 end 17 Output Gbest A. PRE-PROCESSING PHASE (BPSO-DT AND 2D IMAGE CREATION) In this phase, BPSO is applied to implement the feature selection, and the decision tree (DT) [8] is used as BPSO’s fitness function for a classification problem....
[...]
...The first phase is the pre-processing, which included the optimization process using binary particle swarm optimization with design trees (BPSO-DT) algorithm to select the best features of RNA-Seq then converted it to 2D images....
[...]
54 citations
48 citations
Cites methods from "Analysis of Various Decision Tree A..."
...A CART decision tree was used as it can provide accurate results even with a small number of demonstrations [33] and can easily handle outliers due to different interpretations or variations across multiple demonstrators without overfitting [34]....
[...]
36 citations
References
2,591 citations
"Analysis of Various Decision Tree A..." refers background in this paper
...Gain Ratio differs from information gain, which measures the information with respect to a classification that is acquired based on some partitioning [5]....
[...]
..., all the tuples that fall into a given partition would belong into the same class) [5]....
[...]
...Claude Shannon studied the value or “information content” of messages and gave information gain as a measure in his Information Theory [5]....
[...]
...A constraint is added to avoid such condition, whereby the information gain of test selected must be large- at least as great as the average gain over all tests examined [5]....
[...]
436 citations
287 citations
160 citations
63 citations
"Analysis of Various Decision Tree A..." refers background or methods in this paper
...These steps are simple and fast and thus Decision Tree can be applied to any domain [10]....
[...]
...To construct a decision tree, ID3 uses a top-down, greedy search through the given sets, where each attribute at every tree node is tested to select the attribute that is best for classification of a given set [10]....
[...]