Open Hub
Open Hub

 Settings |  Report Duplicate

0
I Use This!
Very Low Activity

Commits : Listings

Analyzed about 4 hours ago. based on code collected 1 day ago.
Nov 05, 2025 — Nov 05, 2026
Commit Message Contributor Files Modified Lines Added Lines Removed Code Location Date
--------------------------------------------------------------------------------------------- DeepLearning BackPropagation Implementation Update --------------------------------------------------------------------------------------------- *) DeepLearning BackPropagation code has been changed to have 3 inputs, 3 hidden and 3 output layers with 3*3=9 input-hidden weights and 3*3=9 hidden-output weights with total of 18 weights *) Software Analytics example has been updated with a third input and logs for it have been committed to testlogs/. *) Logs include a diff notes of how to extend this for arbitrary inputs and accuracy of backpropagation which beautifully converges at ~10^-26 error after ~1000000 iterations.
Srinivasan Kannan
More... almost 10 years ago
Updated code comments
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
------------------------------------------------------------------------------------------------------ DeepLearning Convolution Network - an example pattern recognition from bitmap images ------------------------------------------------------------------------------------------------------ *) Few more bitmap images have been included with same pattern but of different sizes *) There are 5 patterns and 9 bitmaps: - Thin X and Boldfaced X - Thin 0 and Boldfaced 0 - Thin 8 and Boldfaced 8 - No pattern - Thin 1 and Boldfaced 1 *) Expectation is that Convolution Network final neuron layer must output similar values for same pattern and different values for different patterns *) pooling_neuron_weight has been reintroduced in final neuron layer of 10 neurons each with a randomly chosen weight. *) Logs show for all 10 neurons, final neural activation values are close enough for similar patterns and different for different patterns. *) Thus Convolution Network which is a recent advance in DeepLearning works quite well to recognize similar image patterns.
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
Some changes done to Convolution computation: *) Made sigmoid perceptron optional in final neural network from Maxpooling layer so that weighted sum is printed instead of sigmoid value *) This causes the convolution network to be very sensitive to presence of pattern *) 2 more example bitmaps have been added with varying degree of pattern prominence (a pronounced X and a thin 1) *) With this final neurons from maxpooling layer print values of neurons which quite closely reflect the pattern's magnitude: Huge pattern results in big value while small pattern results in small value proportionately. *) This can rank the bitmaps in increasing degree of magnitude of pattern presence *) Logs for this have been committed to testlogs/
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
PythonSpark+Cython Cloud Implementation of Interview Algorithm has been updated to print average clustering coefficient of the definition graph for text. Average clustering coefficient is average(per node clustering coefficient) for all nodes where clustering coefficient of a vertex is a ratio of edges to neighbours to all possible edges to neighbours (or) amount of "cliqueness" of each neighbourhood of all vertices in a graph.
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
---------------------------------------------------------------------------------------- More detailed shopping cart example 2 for RGO+ThoughtNet based Recommender System ---------------------------------------------------------------------------------------- *) python-src/DeepLearning_ReinforcementLearningRecommenderSystems.py has been updated to choose top percentile classes of text input so that vertices with large core numbers get more weightage. *) New shoppingcart2 product reviews text has been added and corresponding Hypergraph history has been created. This example has detailed product description with an assorted mix (TV reviews, Washing Machine reviews, Home Theatre reviews etc.,) - python-src/RecommenderSystems/RecommenderSystems_Edges.shoppingcart2.txt, python-src/RecommenderSystems/RecommenderSystems_Hypergraph_Generated.shoppingcart2.txt *) New input file python-src/RecommenderSystems.shoppingcart2.input.txt has been added with an example product (TV). *) Recommender System chooses relevant TV products from Hypergraph history and displays to user. *) Logs have been committed to testlogs/
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
------------------------------------------------------------------------------------------------------------------------------------- An example usecase of Recommender Systems for online shopping cart ------------------------------------------------------------------------------------------------------------------------------------- *) RecommenderSystems folder has been updated with new text file from an example past shopping cart history items (e.g books from miscellaneous topics in Amazon online bookstore) *) From the above edges a Hypergraph is created by classifying the shopping cart items into classes and constructing hyperedges across the classes (by Recursive Gloss Overlap Graph maximum core number classifier) *) Text files for above usecase have .shoppingcart suffixes *) Above shopping cart hypergraph has a past history of items chosen by a user from variety of topics. *) New input file RecommenderSystems.shoppingcart.input.txt has present items in user's shopping cart and new items have to be recommended to user based on present choice and past history. *) For this, items in present shopping cart are lookedup in past sales history Hypergraph created previously and recommendations are returned *) Items in present shopping cart are looked up in two ways: 1) By classifying with Recursive Gloss Overlap graph and looking up rare classes less than a threshold (presently core number 5 and 2) Raw token lookup *) Logs for these two lookups based Recommendations generated are committed to testlogs/
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago
Kafka Streaming Platform has been added as a data storage in Streaming Generator Abstraction Single-Window entrypoint. Kafka subscriber code for a neuronraindata topic polls for incoming messages in iterator. As an example Streaming CountMeanMinSketch implementation has been updated to source streamed data published in Kafka.
Srinivasan Kannan
More... almost 10 years ago
Updated AsFer Design Document
Srinivasan Kannan
More... almost 10 years ago