thnkrSDK 0.0.7

Software Development Kit for connecting to the thnkrAPI

Description

Welcome to thnkrAI’s Pricing Model and SDK documentation.

thnkrAI allows you to conduct detailed pricing analysis on any kind of product or product datasets you want in an easy and accessible way. thnkrAI’s pricing model can bring enabling price predictions, discovery, and insights to your data.

Use the thnkrAI SDK to quickly read product information and predict prices using the model’s price prediction endpoint.

Model Overview

thnkrAI’s General Pricing Model

Model Description: Text-to-Price Model takes in product data from everyday consumer products and predicts an estimated price based on that information.

Model Type: Natural Language Processing (NLP); Regression Model

Model Input: String of product data (title/description)

Model Output: Numerical float as the predicted price ($USD)

Key Features: Natural Language Processing, Character vectorization, Regression.

Model Evaluation: Evaluated using Mean Absolute Percentage Error (MAPE), R-Squared (R2), and Mean Squared Error (MSE).

Installation

!pip install thnkrSDK

thnkrSDK Integration

LICENSE: MIT

import thnkrSDK as thnkr

model = 'general_v1'

data = {
    "Title": "EXAMPLE PRODUCT TITLE/DATA",
}

username = "USERNAME"
password = "PASSWORD"

prediction = thnkr.predict(model, data, username, password)

Example Output:

{'Prediction': 59.99}

Contact Information

• Contact the thnkrAI team at [e-mail] with any questions or concerns

Model Limitations

• Model is currently trained on limited text data from products listed online. The products come from one online ecommerce marketplace, meaning products found on other marketplaces may be interpreted differently by the model. We are currently working on collecting more product data from a multitude of various online sources to combat this issue.
• If the product data input is not descriptive enough or lacks substantial detail, the model may return inaccurate price predictions compared to those with substantial detail. To combat this, we are working on training the model with more diverse lengths of input strings.

Support

• If you have any issues or questions please contact us at [e-mail]

Version History

• Trained November 15, 2023: thnkrAI general_v1

Base URL

• https://api.thnkrai.com

Parameters

thnkr.predict('model', 'data_input', 'username', 'password')

Errors

This API uses the following error codes:

• 200 Successful Request: The request was malformed or missing required parameters.
• 400 Bad Request: The request was malformed or missing required parameters.
• 401 Unauthorized: The API key provided was invalid or missing.
• 404 Not Found: The requested resource was not found.
• 500 Internal Server Error: An unexpected error occurred on the server.