Question & Answer

Introduction

When users type certain questions into Google, like "What's a recipe for chocolate chip cookies?", they can get an answer directly on the search results page. This answer highlights a piece of content from elsewhere, without requiring the user to exit and go to another site. It's a strategy that's been very effective, helping users get information more quickly and keeping them engaged without leaving the Google search results page.

Miso semantically analyzes and understands your content at a deep level and can also decipher the precise intent behind a user's question. The result is a highly powerful Q&A API that lets you mirror Google's functionality by answering users' questions with paragraphs extracted from the html field for your products. For example, Miso can take a question like What is python?, and extract an answer such as Python is an interpreted, object-oriented, high-level programming language. The raw text that Miso generates answers against is derived from a product's html field.

Before You Start

API Call / JSON Response

To fetch your Q&A answers, POST to the Q&A API with the following minimum parameters:

• q: The question user has entered.
• min_probability: Miso assigns each answer a probability score (between 0.0 and 1.0) that determines how likely a paragraph can accurately answer the question. Answers that score lower than min_probability will be excluded from the response. A probability at least 0.7 is recommended, but you usually will need to fine-tune this threshold to find the precision-and-recall sweet spot for your application.

The API will return a list of nested list of potential answers, ranked by their probability score.

Example API Call

POST /v1/qa/question_answering?api_key=<API_KEY>
{
    "q": "what is gradient descent",
    "min_probability": 0.90,
    "rows": 5,
    "fl": ["title", "authors"]
}

Example JSON Response

{
   "message": "success",
   "data": {
      "took": 153,
      "miso_id": "937c5d70-ddf7-11eb-832d-ba7ee8b57635",
      "total": 38,
      "spellcheck": {
         "spelling_errors": false,
         "auto_spelling_correction": false,
         "original_query": "what is gradient descent",
         "original_query_with_markups": "what is gradient descent",
         "corrected_query": "what is gradient descent",
         "corrected_query_with_markups": "what is gradient descent"
      },
      "answers": [
         {
            "product_id": "9781492037330_ch02.html",
            "answer": {
               "probability": 0.99267578125,
               "text": "Gradient descent is a member of this path-finding class of algorithms. Variations of gradient descent exist, but ...",
               "css_selector": ":root>section>div>section:nth-of-type(2)"
            },
            "title": "Fundamentals of Deep Networks",
            "authors": [
               "Adam Gibson",
               "Josh Patterson"
            ]
         }
        ...
      ]
   }
}

Before You Launch