{"mode":"normal","description":"

Turn card to reveal more information on each algorithm.<\/p>\n","dialogs":[{"tips":{},"text":"

DECISION TREE<\/strong><\/p>\n\n

A set of rules to partition a data set based on the values of the different predictors.<\/p>\n","answer":"

Use cases: <\/strong>Marketing segmentation, fraud detection.<\/p>\n\n

(+) Intuitive to explain to nontechnical business users.<\/p>\n\n

(+) Normalizing predictors is not necessary<\/p>\n\n

(- )Tends to overfit the data.<\/p>\n\n

(-) Selecting the right parameters can be challenging.<\/p>\n\n

(-) Small changes in input data can yield substantially different trees.<\/p>\n"},{"tips":{},"text":"

K-NEAREST NEIGHBOURS.<\/strong><\/p>\n\n

A lazy learner where no model is generalized. Any new unknown data point is compared against similar known data points in the training set.<\/p>\n\n

<\/p>\n","answer":"

Use case:<\/strong> Image processing, applications where slower response time is acceptable.<\/p>\n\n

(+) Requires very little time to build the model.<\/p>\n\n

(+) Handles missing attributes in the unknown record gracefully.<\/p>\n\n

(+) Works with nonlinear relationships.<\/p>\n\n

(-) The deployment runtime and storage requirements will be expensive.<\/p>\n\n

(-) Arbitrary selection is acceptable of value of k.<\/p>\n\n

(-) No description of the model.<\/p>\n"},{"tips":{},"text":"

NA\u00cfVE BAYESIAN<\/strong><\/p>\n\n

Predicts the output class based on the Bayes\u2019 theorem by calculating class conditional probability and prior probability.<\/p>\n","answer":"

Use case:<\/strong> Spam detections, text mining. <\/p>\n\n

(+) Attributes need to be independent.<\/p>\n\n

(+) Training data set needs to be representative sample of population and needs to have complete combinations of input and output. <\/p>\n\n

(-) Time required to model and deploy is minimum. <\/p>\n\n

(-) Great algorithm for benchmarking.<\/p>\n\n

(-) Strong statistical foundation. <\/p>\n"},{"tips":{},"text":"

ARTIFICIAL NEURAL NETWORKS.<\/strong><\/p>\n\n

A computational and mathematical model inspired by the biological nervous system. The weights in the network learn to reduce the error between actual and prediction.<\/p>\n","answer":"

Use cases: <\/strong>Image recognition, fraud detection, quick response time applications.<\/p>\n\n

(+) Good at modeling nonlinear relationships.<\/p>\n\n

(+) Fast response time in deployment<\/p>\n\n

(-) No easy way to explain the inner working of the model.<\/p>\n\n

(-) Requires preprocessing data. <\/p>\n\n

(-) Cannot handle missing attributes.<\/p>\n\n

<\/p>\n"},{"tips":{},"text":"

K-MEANS.<\/strong><\/p>\n\n

Data set is divided into k clusters by finding k centroids.<\/p>\n\n

<\/p>\n","answer":"

Use case:<\/strong> Customer segmentation, anomaly detection, applications where globular clustering is natural.<\/p>\n\n

(+) Simple to implement<\/p>\n\n

(+) Can be used for dimension reduction<\/p>\n\n

(-) Specification of k is arbitrary and may not find natural clusters.<\/p>\n\n

(-) Sensitive to outliers<\/p>\n\n

<\/p>\n"},{"tips":{},"text":"

SELF-ORGANISING MAPS.<\/strong><\/p>\n\n

A visual clustering technique with roots from neural networks and prototype clustering.<\/p>\n\n

<\/p>\n","answer":"

Use case:<\/strong> Diverse applications including visual data exploration, content suggestions, and dimension reduction.<\/p>\n\n

(+) A visual way to explain the clusters.<\/p>\n\n

(+) Reduces multidimensional data to two dimensions.<\/p>\n\n

(-) Number of centroids (topology) is specified by the user.<\/p>\n\n

(-) Does not find natural clusters in the data.<\/p>\n"}],"behaviour":{"enableRetry":true,"disableBackwardsNavigation":false,"scaleTextNotCard":false,"randomCards":false,"maxProficiency":5,"quickProgression":false},"answer":"Turn","next":"Next","prev":"Previous","retry":"Retry","correctAnswer":"I got it right!","incorrectAnswer":"I got it wrong","round":"Round @round","cardsLeft":"Cards left: @number","nextRound":"Proceed to round @round","startOver":"Start over","showSummary":"Next","summary":"Summary","summaryCardsRight":"Cards you got right:","summaryCardsWrong":"Cards you got wrong:","summaryCardsNotShown":"Cards in pool not shown:","summaryOverallScore":"Overall Score","summaryCardsCompleted":"Cards you have completed learning:","summaryCompletedRounds":"Completed rounds:","summaryAllDone":"Well done! You got all @cards cards correct @max times in a row!","progressText":"Card @card of @total","cardFrontLabel":"Card front","cardBackLabel":"Card back","tipButtonLabel":"Show tip","audioNotSupported":"Your browser does not support this audio","confirmStartingOver":{"header":"Start over?","body":"All progress will be lost. Are you sure you want to start over?","cancelLabel":"Cancel","confirmLabel":"Start over"},"title":""}

node-5030__activity-4-temporary