{"repo":"mostly-ai/virtualdatalab","free":true,"listed":false,"github":"https://github.com/mostly-ai/virtualdatalab","clone":"git clone https://github.com/mostly-ai/virtualdatalab.git","description":"Benchmarking synthetic data generators for sequential data in terms of accuracy and privacy.","language":"Python","stars":21,"topics":["synthetic-data","sequential-data","privacy"],"license":"GPL-3.0","category":"self-hosted-apps","readme_excerpt":"UPDATE : Please go to https://github.com/mostly-ai/mostlyai-qa for the open-source Synthetic Data Quality Assurance toolkit. ---- Virtual Data Lab (VDL) The Virtual Data Lab is a python-based framework to assess generative models for sequential data with respect to their accuracy as well as privacy given a range of real-world and machine-generated datasets. In addition, it contains basic synthesizers capable of sequential data generation. The DL-based generative models, developed by MOSTLY AI as part of their commercial offering, are not part of this package. License: GPLv3 It is comparable to the SDGym by MIT, but differs in terms of its sole focus on sequential data, and its inclusion of privacy metrics. Installation It is recommended to install virtualdatalab in a Conda virtual environment to avoid interfering with other software programs. Data Requirements Synthesizers and public functions accept only data formatted according to the following guideline. The generated data will adhere to that same format. common data format : data resides within a single pandas DataFrame DataFrame indexes: id - identifying a subject/person/indivudal sequence pos - order of sequences belonging to one subject DataFrame columns are either of type numeric or categorical No NaN values. NaN values in categorical columns will be treated as the empty string The helper method target data manipulation.prepare common data format can read in CSV files, and provides data in the required format. Feature","default_branch":null,"files":null,"tree":[],"storefront":"/r/mostly-ai","claimed":false,"request_supported":{"post":"https://gitbuyer.com/r/mostly-ai/virtualdatalab/request-supported","requests":0},"note":"indexed from public GitHub; nothing is for sale on this page. Clone it from GitHub. Paid listings live at /search."}