Workflows

Creation of New Attribute Depending on Val... (1)

• Download

The process shows the usage of the operator "Generate Attributes" in combination with an "if - then - else" condition and nominal values. The values "value0" and "value1" are mapped to "T1", other values are mapped to "T2" for the new attribute.

Created: 2010-06-01

Association rules as examples (1)

• Download

Uses Groovy scripting to lay out rules and their support metrics as examples.

Created: 2010-06-02

Combining nominal attributes with missing (1)

• Download

This process will show how one can combine nominal values of attributes that contain missing values. A generate attribute operator is used and hence forbidden characters must first be replaced. After this, a condition in the generation ensures that no question mark (standing for missing value) will be shown in the result, if any of the two combined attributes is known.

Created: 2010-06-04

Prepending common prefix to attributes (1)

• Download

This process shows how one can add a common prefix to a subset of attributes. First a subset might be defined by the attribute set selection parameters of the rename by replacing operator. Then one can make use of the capturing group functionality of regular expressions to insert the original name into the replacement.

Created: 2010-06-04

Transform Attribute Names to Upper Case (S... (1)

• Download

This process uses a Script operator which transforms the attribute names of the input example set into UPPER case.

Created: 2010-06-06

Transform Attribute Names to lower Case (S... (1)

• Download

This process uses a Script operator which transforms the attribute names of the input example set into lower case.

Created: 2010-06-06

Same Number of Examples per Class (Stratif... (1)

• Download

This process can be used to sample examples from each class of the data set so that the number of examples per class is the same for all classes. The name of the label attribute is defined in the first "Set Macro" operator within the subprocess "Stratification". The result will be a stratified data set where each class is represented by the minimum number of examples for a single class minus 1 (due to calculation reasons in absolute sampling which is used here). The first two operators just...

Created: 2010-06-10

Function finding (1)

• Download

This process recursively generates new attributes by mathematically combining existing attributes with user selectable operators. At the end of each pass through the recursion attributes are thinned to prevent silicon smoke and unpleasantness. Sample input. P1    P2    A1    A2    RESULT 2       3      4      5      15 6     &...

Created: 2010-06-12 | Last updated: 2010-06-29

Preprocessing nominal data for frequent it... (1)

• Download

This process will first create artificial data that can be compared to usual data loaded for frequent item set mining: Nominal Data with a true and false value, but differently mapped to internal indices. For ItemSet Mining these must be preprocessed to avoid problems: First they have to be transformed to Binominal Attributes, then it has to be defined, which is the positive and the negative value.

Created: 2010-06-16 | Last updated: 2010-06-16

Optimizing Discretization (1)

• Download

This process generates a decision tree on the Iris data set. Before the decision tree is generated, the input attributes are discretised so we only work on nominal attributes. We use a combination of "Select Subprocess" and "Optimize Parameters" to select the best out of five different discretizazion methods independently for each of the attributes. The process shows, that the resulting accuracy heavily depends on the choice of the method. It varies between 64% and 94%.

Created: 2010-06-18 | Last updated: 2010-06-18

Finding all Examples that have duplicate v... (1)

• Download

This process will retrieve all examples, who have identical values in a specific attribute. For testing, the following data can be writen into the file, that will be read by the Read CSV operator: CID,Value 3596,X 4054,X 4054,X 3000,S 3000,T 3000,U 32135,S The target of this process is to return the two examples having the same value in the CID column. To achieve this, first a real id is generated by the generate id. After this, we have to find all duplicates: For this we first remove dupl...

Created: 2010-06-18

Discretization into Deviation Interval aro... (1)

• Download

This process shows how numerical attributes can be discretized into intervals based on the standard deviation for each attribute around their mean values.

Created: 2010-06-22 | Last updated: 2010-06-22

Pre process data per group using recall re... (1)

• Download

This process takes as input a dataset with different groups. For each group it discretises attributes into High, Medium, Low. This creates a group invariant representation.

Created: 2010-06-23 | Last updated: 2010-06-23

Macro Propagation into Subprocesses and Ex... (1)

• Download

<p> This small process demonstrates how to propagate macros into a subprocess or Execute Process operator </p> <p> When macros are set once, they are available in subprocesses imediatly (see Generate Attributes in the seconde Subprocess operator) </p> <p> If someone wants to propagate a macro into a Excecute Process operator one need to edit the "macros" parameter in the Parameters list of the Execute Process operator. </p> Tags: Subprocess, Execute Proc...

Created: 2010-06-25

Holograph Reduced Representation (1)

• Download

After Kanerwa

Created: 2010-06-29

Generating Example Weights depending on label (1)

• Download

This process uses a Generate Attribute operator for assigning new weights to examples. It uses the if condition of this operator to distinguish between labels. This can be especially useful if you have a strong bias in your class frequency, which can harm learning. Please note that you must use a learning algorithm that supports weighted examples.

Created: 2010-05-03

Weighted Score Tables (1)

• Download

This process calculates a weighted score as known from SAP BW. The first operator generates data similar to that used in the links you provided. The next operator (Discretize) defines the different age groups. The next three operators map each age group to the value used for scoring. The last operator finally calculates the score and adds it as a new column to your data set.

Created: 2010-05-05

Simple Recommender System Web Service (1)

• Download

 This workflow takes macro value of user request and computes top-n recommendations from already learned model on train set. Macro value is transformed to example set with the appropriate \textit{user identification} role within \textit{ProcessInput} subprocess operator.   Simple recommendation web service workflow takes three inputs: \begin{enumerate} \item Macro value \textit{user} - identification number of user request \item Train set from RM repository \item Learned Mo...

Created: 2012-05-10 | Last updated: 2012-05-10

Calibration for execution time extension (1)

• Download

 Benchmark workflow that loops Naive Bayes operator many times and provides estimate of the system's overall performance.

Created: 2012-05-10 | Last updated: 2012-05-10

Front-end Recommendation Web Service (1)

• Download

 Front-end recommendation web service has a simple job to query the cached recommendations from the \textit{item recommendation table}

Created: 2012-05-11 | Last updated: 2012-05-11

Write activity web service (1)

• Download

 When the specific user $i$ "consumes" certain item $j$ the write activity web service is called, which writes activity $(i,j)$ to \textit{train set table} and removes recommendation $j$ for user $i$ from \textit{item recommendation table} in SQL database.

Created: 2012-05-11 | Last updated: 2012-05-11

Online update recommendation web service (1)

• Download

 After some arbitrary number of recommendations to specific users, system has to update recommendations in \textit{item recommendation table}. This is accomplished by calling the online update recommendation web service, which updates the recommendation model in RapidAnalytics repository and updates the recommendations for specific users in \textit{item recommendation table}.

Created: 2012-05-11 | Last updated: 2012-05-11

Offline update recommendation web service (1)

• Download

No description

Created: 2012-05-11

Clustering data from DBpedia using AHC (1)

• Download

 This workflow uses Aglomerative Hierarchical Clustering algorithm to build hierarchy of clusters on data downloaded from DBpedia, the semantic version of Wikipedia. RDF data are downloaded from SPARQL endpoint and merged with DBpedia ontology by "Build Knowledge Base". Set of items to cluster is selected with "SPARQL Selector" and later they are clustered by "Agglomerative Hierarchical Clustering" with distance measure induced by the Bloehdorn Kernel [1]. ...

Created: 2012-05-30 | Last updated: 2012-06-07

Evaluating semantic kernel with k-NN class... (1)

• Download

This workflow uses k-NN classifier to evaluate quality of EL++ Convolution Kernel [1]. As a dataset one of the examples from DL-Learner project [2] is used. After preparing knowledge base with "Build Knowledge Base", the item to item distance matrix is computed with "Calculate Gram/Distance Matrix". Such a matrix is then used as an input to 10-fold cross-validation with k-NN as an classifier and average result is delivered. [1] L. Józefowski, A. Lawrynowicz, J...

Created: 2012-05-30 | Last updated: 2012-06-07

TextualAttributeExtraction for Recommender... (1)

• Download

This workflow takes conent data from VideoLectures.Net Recommender System Challenge and extractes word-vectors for each lecture. Latent semantic analysis with Singular Value Decomposition is done on item-word binary matrix. The last step is the binomializaiton of dataset.

Created: 2012-06-03

Offline update recommendation web service V2 (1)

• Download

  Periodically we have to do a full re-training on whole train set by the offline update recommendation web service.

Created: 2012-06-20 | Last updated: 2012-06-20

Online update recommendation web service V2 (1)

• Download

 After some arbitrary number of recommendations to specific users, system has to update recommendations in item recommendation table. This is accomplished by calling the online update recommendation web service, which updates the recommendation model in RapidAnalytics repository and updates the recommendations for specific users in item recommendation table.

Created: 2012-06-20 | Last updated: 2012-06-20

Simple Recommender System Web Service V2 (1)

• Download

No description

Created: 2012-06-20

Instance Selection and Prototype Based Rul... (1)

• Download

Example of Instance selection (Requires Instance Selection and Prototype Based Rules plugin): Before training kNN we do instance selection based on ENN algorithm. The ENN algorithm can be replaced by any other instance selection algorithms from Prules/Selection/* Instance selection algorithms form PRules plugin works as View on original ExampleSet

Created: 2011-11-05 | Last updated: 2011-11-05

Instance Selection and Prototype Based Rul... (1)

• Download

Example of Instance selection chain (Requires Instance Selection and Prototype Based Rules plugin): Beafore training kNN we do instance selection based on cascade of instance selection algorithms. Here ENN algorithm is followed by CNN algorithm. Instance selection algorithms form PRules plugin works as View on original ExampleSet

Created: 2011-11-05

Instance Selection and Prototype Based Rul... (1)

• Download

Example of Instance selection: (Requires Instance Selection and Prototype Based Rules plugin): Here instance selection is based on All-kNN algorithm. Because the Pro output of any instance selection operator is just an ExampleSet object, so in fact any Learner can be used instead of kNN, like DecisionTree, but in this example the Decision Tree is based on the reduced ExampleSet .

Created: 2011-11-05

Instance Selection and Prototype Based Rul... (2)

• Download

Example of Instance optimization (Requires Instance Selection and Prototype Based Rules): Here we use FCM (Fuzzy C-means) clustering to initialize kNN classifier. The "Class assigner" is responsible for assigning class label for each cluster center. The class assigner operator use Voronoi diagram and the majority voting for determining class label.

Created: 2011-11-05 | Last updated: 2011-11-07

Instance Selection and Prototype Based Rul... (1)

• Download

Example of Instance optimization - using clustering for training kNN classifier (Requires Instance Selection and Prototype Based Rules) : Here we use FCM (Fuzzy c-means) clustering to initialize kNN classifier. Moreover centers of clusters are determined independent for each class. "Class iterator" operator iterates over each class label, and embedded clustering algorithm cluster the examples for each class independent. The Prototype (Pro) output of "Class iterator" delivers the concatena...

Created: 2011-11-05 | Last updated: 2011-11-05

Instance Selection and Prototype Based Rul... (2)

• Download

Example of Instance optimization using LVQ algorithm for training kNN classifier (Requires Instance Selection and Prototype Based Rules): Here we use FCM clustering to initialize LVQ network. The "Class assigner" is responsible for assigning class label for each cluster center, then obtained ExampleSet is used as codebooks initialization in the LVQ operator, which on the Prototype (Pro) output delivers the new optimized position of codebooks (prototypes) for training nearest n...

Created: 2011-11-05 | Last updated: 2011-11-07

Series of Factorials (1)

• Download

 The workflow generates the series of factorial numbers.

Created: 2011-11-09 | Last updated: 2011-11-11

Handling data and time example (1)

• Download

This process shows how date and time formats can be converted from arbitrary formats to other arbitrary formats with the operators "Nominal to Date" and "Date to Nominal".

Created: 2011-11-28

Parameter optimization (1)

• Download

This is a parameter optimization workflow for rating prediction recommendation operators.

Created: 2012-02-10 | Last updated: 2012-02-10

Credits: Matej Mihelčić

De-normalizing Dataset After K-means (1)

• Download

K-means is performed on the iris dataset. After finding the clusters, the dataset is de-normalized and the means of the different variables are calculated by cluster.

Created: 2011-10-19 | Last updated: 2011-10-19

POSTing CSV file to RapidAnalytics Web ser... (1)

• Download

This is the first demo used in the RapidAnalytics video on creating Web services. It downloads three data sets provided by data.gov.uk and generates a regression model and stores it as "RegressionTree" in the repository.

Created: 2011-11-02

POSTing CSV file to RapidAnalytics Web ser... (1)

• Download

This is the second demo process used in the RapidAnalytics video on creating Web services. This process is the actual scoring process and used the model generated by the first process. The first input must be a CSV blob in the repository. Once the process is exposed as a Web service in RapidAnalytics, this input will be replaced by the body of the HTTP POST request.

Created: 2011-11-02 | Last updated: 2011-11-02

RCOMM 2011 Challenge 1: Hobbit Genealogy (1)

• Download

This is a solution for Challenge 1 of the a live data mining process design competition "Who Wants to be a Data Miner" held at RCOMM 2011 in Dublin. As you certainly know, Balbo Baggins is the common ancestor of Balbo and Frodo Baggins. The file opened by the operator "Open Ancestor" contains a table with details about parentship in the Baggins family (insert a breakpoint after read CSV). Each example contains a parent and a child. Of course, the same parent can be contai...

Created: 2011-11-02 | Last updated: 2011-11-02

RCOMM 2011 Challenge 3: RapidDraw (1)

• Download

This is a solution for Challenge 3 of the a live data mining process design competition "Who Wants to be a Data Miner" held at RCOMM 2011 in Dublin. The task was to generate a dataset that looks like a spiral when viewed in an appropriate plotter. This process opens a file with three initial data points and subsequently adds more points to the data set in a loop, using macros to extract data values of the predecessors and a "Generate Attributes" operator to add new data points. To view the...

Created: 2011-11-02

weather (1)

• Download

No description

Created: 2011-12-15

weather (1)

• Download

No description

Created: 2011-12-15

Item recommendation collaborative-based wo... (1)

• Download

This workflow takes input 2 as a train set for item-k-NN recommender system. The model peformance is evaluated on test set (input 1). This workflow uses recommender system extension. Train and test set must contain user_id and item_id atrributes which need to have special roles user identification and item identification.

Created: 2012-01-09 | Last updated: 2012-01-09

Item recommendation hybrid-based workflow (1)

• Download

This workflow takes input 2 as a train set for recommender systems. We build two item recommendation models: item k-NN (collaborative based) and item attribute k-NN (content based). Item attribute k-NN operator takes additional item attributes from input 3. We combine two models with operator model combiner and test performance on test set (input 1). Train and test set must contain user_id and item_id attributes which need to have special roles user identification and item identification. Th...

Created: 2012-01-19

Item rating predictions collaborative-based (1)

• Download

This workflow takes input 2 as as train set for several item recommendation predictions. We build four different recommendation model which are combined into one model with operator Model Combiner. Then we take input 1 as a test set and apply merged model. Train and test set must contain user_id, item_id and rating attributes which need to have special roles user identification, item identification and label. This workflow uses recommender system extension.

Created: 2012-01-19

Format_transform_RMonto-RSextension (1)

• Download

This workflow provides transformation of an item description attribute set from RMonto operator into a format required by attribute based k-NN operators of the Recommender extension.

Created: 2012-03-26

MyExperimentTest (1)

• Download

No description

Created: 2013-03-20