Perform Machine Learning Classification


What is Weka

Named after a flightless New Zealand bird, Weka is a set of machine learning algorithms that can be applied to a data set directly, or called from your own Java code. Weka contains tools for data pre-processing, classification, regression, clustering, association rules, and visualization.



Classification

 In machine learning and statistics, classification is the problem of identifying to which of a set of categories (sub-populations) a new observation belongs, on the basis of a training set of data containing observations (or instances) whose category membership is known. An example would be assigning a given email into "spam" or "non-spam" classes or assigning a diagnosis to a given patient as described by observed characteristics of the patient (gender, blood pressure, presence or absence of certain symptoms, etc.). Classification is an example of pattern recognition [wikipedia].

Steps to perform classification using Weka

  In this example we have two data sets, i.e. 'training_data' , which is used for training the model and 'test_data', which is used for actual classification of the flood traffic. Following are the steps that can be followed in order to perform classification of any sort of training and testing data.

 1. Save both data sets into csv files to be opened in Weka
2. Open both data sets into Weka and classify using zero-R and click on use training set and run the ZeroR algorithm
3. Click Start to run algorithm and in the status window and click on visualize classifier errors and save the data.arff and predict.arff file
4. Open training_set.arff and training_set.csv in notepad++ and copy the dataset from csv file and paste into data.arff file
5. Delete @predict attribute from data.arff file
6. Open test_set.csv and test_set.arff file into notepad++ and copy the dataset from csv file into arff file
7. Open training_set.arff file and test_set.arff file in notepad++ and copy attributes from training_set.arff file and paste into test_set.arff file under @relation
8. Put question mark at the end of each data set relation in test_set.arff file
9. Open training_set.arff in Weka file and click on classify and choose multilayer regression and choose cross validation 10 fold and click on start now click on save model and save it with the name model
10. Open test_set.arff file and click on classify and choose multilayer regression and load model and choose the model saved
11. Click on supplied test set and choose predict.arff file
12. Click on revaluate model using supplied test set, now open the test_set.arff file in Weka

Comments

Post a Comment

Popular posts from this blog

Guidelines for Effective Academic Writing

 Here are some guidelines for effective academic writing. It is based on my own experience writing for computer science journals and conferences.  Don't use 'the' with an abbreviation. Use active tenses unless a passive is extremely necessary. Use the same form of the verb (present/past) when describing related work and be consistent with it throughout the manuscript. Don't use the before plurals. Don't use etc when including a list of items. Never use the word 'suggest' when referring to the previous work. The word 'report' could be a better option.  Don't use 's with an abbreviation, for example, SDN's capability could be written as SDN capability. Don't use the word "paradigm" use only when it makes sense. Always use "," after then, however, thus, etc. Be consistent, use the same form of "TENSE, i.e., PAST or PRESENT" in the whole section. Do not use "the", "a" ...in front of single
Read more

Unstructued Notes on TCP IP Networking

Image
 OSI:- theoretical model, networking, how data packets are created and processed, used by TCP/IP. OSI open systems interconnection model.  Physical layer: lowest, NIC, RJ45, physical medium data link layer:-data packet is converted into bits, mac address resides network layer:-One to another. routing, and logical addressing, IPv6, IPv4, error handling Transport Layer:-TCP and UDP exist establish connectivity between two machines Session Layer:-Establishes and terminates connection among computers, control session TCP/IP sessions presentation layer:-Data formatting, description,  application layer:- application uses the protocol to use the data Unicasting, multicasting, broadcasting:- unicast from single user to another single user, one user multi-users, broadcast broadcast  DNS:-Domain name system, cybersecurity.com--->it will be converted this address to the IP address done by the DNS server What us firewall:-Hardware or software which allows or black traffic.  Packet filtering pac
Read more

Protecting IT Infrastructure: Key Takeaways from the CrowdStrike Update Incident

Image
  The Microsoft outage, linked to CrowdStrike’s Falcon Sensor, led to 'Blue Screen of Death' errors and widespread service disruptions in various industries, including hospitals, banks, airlines, emergency services, and supermarkets. In the recent incident, an update to CrowdStrike's Falcon product caused significant disruptions globally. Falcon is an endpoint detection and response (EDR) solution that monitors traffic passing through systems to protect against malicious files, viruses, and malware, relying on cloud technology to secure devices on corporate networks (CrowdStrike, 2024). The problem arose when an update to the Falcon software triggered an endless loop of the Blue Screen of Death (BSOD) on many Windows machines (Microsoft, 2024). This catastrophic failure occurred because the Falcon software talks directly to the Windows kernel for greater speed and protection. The update contained a bug that led to a kernel-level error, causing the machines to crash and beco
Read more