search

alan-turing-institute / QUIPP-collab

Collaboration on the QUIPP project
1 stars 1 forks source link

US adult census dataset #29

Open kasra-hosseini opened 5 years ago

kasra-hosseini commented 5 years ago

Link: http://archive.ics.uci.edu/ml/machine-learning-databases/adult/

The dataset has 30,162 records and a binomial label indicating a salary of less than $50,000 or greater than $50,000. Of the records in the dataset, 75% have a class label of greater than $50,000. There are 14 attributes consisting of eight categorical and six continuous attributes.

kasra-hosseini commented 5 years ago
Screenshot 2019-11-15 at 09 55 29
kasra-hosseini commented 5 years ago

Some info about the dataset (from this blog):

The dataset used in this project has 30,162 records and a binomial label indicating a salary of less than $50,000 or greater than $50,000. Of the records in the dataset, 75% have a class label of greater than $50,000. There are 14 attributes consisting of eight categorical and six continuous attributes (Table 1). The employment class describes the type of employer, such as self-employed or federal, and occupation describes the employment type, such as farming, clerical or managerial. Education contains the highest level of education attained, such as high school or doctorate. The relationship attribute has categories such as unmarried or husband, and marital status has categories such as married or separated. The other nominal attributes are country of residence, gender and race. The continuous attributes are age, hours worked per week, education number (numeric representation of the education attribute), capital gain and loss, and a weight attribute that is a demographic score assigned to an individual based on information such as state of residence and type of employment. The continuous variable “fnlwgt” represents final weight, which is the number of units in the target population that the responding unit represents. It is used to weight up the survey results and it is not intrinsic to the population, but an external variable that is dependent on the sample design. However, for the sake of the present analysis we do treat this variable as a “normal” feature and train our GANs on the complete dataset. As we will show later, absence of correlations between this feature and the rest of the features will be evident in the results obtained.

martintoreilly commented 5 years ago

@LouiseABowler @kasra-hosseini Can this and issue #15 be merged?

kasra-hosseini commented 5 years ago

@LouiseABowler in issue #15 wrote:

US adult census dataset. Used in the ONS blog post as an example. Can be used in a classifier task - are people earning above or below $50,000 based on other characteristics?

https://archive.ics.uci.edu/ml/datasets/adult