ML Categorize
This section outlines all of the elements that have to do with running the ML Categorize functionality on the Sincera application via Private Jobs.
Overview
The ML Categorize functionality takes and returns the “Collect Text” module details and leverages Sincera’s ML engine to provide a detailed weighted categorization for a given URL using the IAB 3.0 taxonomy.
The ML Categorize private job type will return the same details as the “Collect Text” module but will also contain two additional columns. These columns are cat_result and rich_payload.
Category Results (cat_results)
Specifications
| Field | Data Type | Definition |
|---|---|---|
| IAB3.0[*] | STRING | The name of the IAB category identified on the page |
| IAB3.0[*].score | STRING | The model’s confidence that the category is on the page |
| IAB3.0[*].tier-2[*] | STRING | The name of the IAB subcategory identified on the page |
| IAB3.0[*].tier-2[*].score | STRING | The model’s confidence that the subcategory is on the page |
| IAB3.0[*].tier-2[*].tier-3[*] | STRING | The name of the IAB tier 2 subcategory identified on the page |
| IAB3.0[*].tier-2[*].tier-3[*].score | STRING | The model’s confidence that the tier 2 subcategory is on the page |
| IAB3.0[*].tier-2[*].tier-3[*].tier-4[*] | STRING | The name of the IAB tier 3 subcategory identified on the page |
| IAB3.0[*].tier-2[*].tier-3[*].tier-4[*].score | STRING | The model’s confidence that the tier 3 subcategory is on the page |
Example Output
{
"IAB3.0": {
"pop culture": {
"score": "0.106508076",
"tier-2": {
"celebrity style": {
"score": "1.00"
}
},
},
"genres": {
"score": "0.10101056"
},
"personal celebrations & life events": {
"score": "0.23227972"
},
"events": {
"score": "0.30574763",
"tier-2": {
"awards shows": {
"score": "1.00"
}
},
},
"style & fashion": {
"score": "0.9704324",
"tier-2": {
"men's fashion": {
"score": "0.94",
"tier-3": {
"men's accessories": {
"score": "0.95",
"tier-4": {
"men's jewelry and watches": {
"score": "0.3456034"
},
},
},
"men's clothing": {
"score": "0.87",
"tier-4": {
"men's formal wear": {
"score": "0.62414813"
},
"men's casual wear": {
"score": "0.13748628"
},
},
},
},
},
},
},
"law": {
"score": "0.28526545"
},
"sensitive topics": {
"score": "0.13958564"
},
"real estate": {
"score": "0.71693695",
"tier-2": {
"apartments": {
"score": "0.31"
}
},
},
"automotive": {
"score": "0.9352126",
"tier-2": {
"auto type": {
"score": "1.00"
},
"auto body styles": {
"score": "1.00",
"tier-3": {
"van": {
"score": "0.99"
},
"convertible": {
"score": "0.39"
},
"off-road vehicles": {
"score": "1.00"
},
},
},
},
},
},
}
Rich Payload (rich_payload)
Specifications
| Field | Data Type | Definition |
|---|---|---|
| sincera-model-name | STRING | The name of the Sincera model being run |
| sincera-model-id | STRING | The Sincera model version used to generate the classifications |
| categorizations | LIST OF OBJECTS | A list of objects for each classification found from the page |
| categorizations[*].score | STRING | The model’s confidence that the category is on the page |
| categorizations[*].unique_id | STRING | An ID that relates to the IAB taxonomy code for the classification |
| categorizations[*].complete_tier | STRING | The top level category name |
| categorizations[*].tier-2 | LIST OF OBJECTS | A list of object for the subcategories identified on that page |
| categorizations[*].tier-2[*].score | STRING | The model’s confidence that the category is on the page |
| categorizations[*].tier-2[*].unique_id | STRING | An ID that relates to the IAB taxonomy code for the classification |
| categorizations[*].tier-2[*].complete_tier | STRING | A combined string of <tier_1_name>/<tier_2_name> |
| categorizations[*].tier-2[*].tier-3 | LIST OF OBJECTS | A list of object for the tier 2 subcategories identified on that page |
| categorizations[*].tier-2[*].tier-3[*].score | STRING | The model’s confidence that the category is on the page |
| categorizations[*].tier-2[*].tier-3[*].unique_id | STRING | An ID that relates to the IAB taxonomy code for the classification |
| categorizations[*].tier-2[*].tier-3[*].complete_tier | STRING | A combined string of <tier_1_name>/<tier_2_name>/<tier_3_name> |
| categorizations[*].tier-2[*].tier-3[*].tier-4 | LIST OF OBJECTS | A list of object for the tier 3 subcategories identified on that page |
| categorizations[*].tier-2[*].tier-3[*].tier-4[*].score | STRING | The model’s confidence that the category is on the page |
| categorizations[*].tier-2[*].tier-3[*].tier-4[*].unique_id | STRING | An ID that relates to the IAB taxonomy code for the classification |
| categorizations[*].tier-2[*].tier-3[*].tier-4[*].complete_tier | STRING | A combined string of <tier_1_name>/<tier_2_name>/<tier_3_name> /<tier_4_name> |
Example Output
{
"sincera-model-name": "Sincera-IAB_3.0",
":sincera-model-id": "10.4",
"categorizations": {
"Pop Culture": {
":score": "0.106508076",
":unique_id": "432",
":complete_tier": "Pop Culture",
"tier-2": {
"Celebrity Style": {
":score": "1.00",
":unique_id": "439",
":complete_tier": "Pop Culture/Celebrity Style",
},
},
},
"Genres": {
":score": "0.10101056",
":unique_id": "SPSHQ5",
":complete_tier": "Genres",
},
"Personal Celebrations & Life Events": {
":score": "0.23227972",
":unique_id": "163",
":complete_tier": "Personal Celebrations & Life Events",
},
"Events": {
":score": "0.30574763",
":unique_id": "8VZQHL",
":complete_tier": "Events",
"tier-2": {
"Awards Shows": {
":score": "1.00",
":unique_id": "162",
":complete_tier": "Events/Awards Shows",
},
},
},
"Style & Fashion": {
":score": "0.9704324",
":unique_id": "552",
":complete_tier": "Style & Fashion",
"tier-2": {
"Men's Fashion": {
":score": "0.94",
":unique_id": "579",
":complete_tier": "Style & Fashion/Men's Fashion",
"tier-3": {
"Men's Accessories": {
":score": "0.95",
":unique_id": "580",
":complete_tier":
"Style & Fashion/Men's Fashion/Men's Accessories",
"tier-4": {
"Men's Jewelry and Watches": {
":score": "0.3456034",
":unique_id": "581",
":complete_tier":
"Style & Fashion/Men's Fashion/Men's Accessories/Men's Jewelry and Watches",
},
},
},
"Men's Clothing": {
":score": "0.87",
":unique_id": "582",
":complete_tier":
"Style & Fashion/Men's Fashion/Men's Clothing",
"tier-4": {
"Men's Formal Wear": {
":score": "0.62414813",
":unique_id": "585",
":complete_tier":
"Style & Fashion/Men's Fashion/Men's Clothing/Men's Formal Wear",
},
"Men's Casual Wear": {
":score": "0.13748628",
":unique_id": "584",
":complete_tier":
"Style & Fashion/Men's Fashion/Men's Clothing/Men's Casual Wear",
},
},
},
},
},
},
},
"Law": {
":score": "0.28526545",
":unique_id": "383",
":complete_tier": "Law",
},
"Sensitive Topics": {
":score": "0.13958564",
":unique_id": "v9i3On",
":complete_tier": "Sensitive Topics",
},
"Real Estate": {
":score": "0.71693695",
":unique_id": "441",
":complete_tier": "Real Estate",
"tier-2": {
"Apartments": {
":score": "0.31",
":unique_id": "442",
":complete_tier": "Real Estate/Apartments",
},
},
},
"Automotive": {
":score": "0.9352126",
":unique_id": "1",
":complete_tier": "Automotive",
"tier-2": {
"Auto Type": {
":score": "1.00",
":unique_id": "16",
":complete_tier": "Automotive/Auto Type",
},
"Auto Body Styles": {
":score": "1.00",
":unique_id": "2",
":complete_tier": "Automotive/Auto Body Styles",
"tier-3": {
"Van": {
":score": "0.99",
":unique_id": "7",
":complete_tier": "Automotive/Auto Body Styles/Van",
},
"Convertible": {
":score": "0.39",
":unique_id": "8",
":complete_tier": "Automotive/Auto Body Styles/Convertible",
},
"Off-Road Vehicles": {
":score": "1.00",
":unique_id": "14",
":complete_tier":
"Automotive/Auto Body Styles/Off-Road Vehicles",
},
},
},
},
},
},
}