¿Todo listo para comenzar?
Ejecute campañas eficaces de UA a partir de aprendizaje automático y targeting contextual.
Calculadora contextual
Al entrar en vigor la retirada del uso de IDFA en iOS 14, los anunciantes se enfrentarán a tener que buscar una solución de targeting alternativa que no sea a nivel de usuario y que siga generando un CPI competitivo y usuarios valiosos. Aunque el juego cambia, el objetivo sigue siendo el mismo: crecimiento eficaz.
A medida que el targeting contextual gana protagonismo, y para obtener información que supere las categorías fácilmente manipulables de la tienda de aplicaciones, usamos un método de representación de palabras conocido como ELMo. Con este, representamos cada agrupación como un vector, lo que permite la medición de la ″distancia″ contextual entre aplicaciones. La distancia calculada se incluye como una de las cientos de otras funciones en nuestros modelos de predicción y tiene un papel importante en nuestra lógica de puja.
Como forma de resaltar su eficacia, creamos la calculadora de distancia contextual. Al usarla, usted obtendrá las aplicaciones más cercanas a la suya en términos de mecánica, tema, funciones y otros aspectos. Se obtendrá mediante la comparación exhaustiva entre los textos de descripción de la tienda de aplicaciones.
Encuentre aplicaciones con relación contextual
Nombre de la aplicación
Categoría
Las principales aplicaciones relacionadas contextualmente
| No | App name | Confidence |
|---|---|---|
| 1 | Billion Cash Slots-Casino Game | 100% Confidence Level 100% - High Confidence As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps. Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity. |
| 2 | Eon Slots Casino Vegas Game | 90% Confidence Level 90% - High Confidence As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps. Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity. |
| 3 | Cash Tornado Slots - Casino | 50% Confidence Level 50% - Moderate Confidence As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps. Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity. |
| 4 | Super Vegas Slots Casino Games | 44% Confidence Level 44% - Moderate Confidence As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps. Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity. |
| 5 | Vegas Nights Slots | 40% Confidence Level 40% - Moderate Confidence As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps. Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity. |
| 6 | Tycoon Casino - Vegas Slots | 90% Confidence Level 90% - High Confidence As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps. Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity. |
| 7 | Winning Slots Las Vegas Casino | 90% Confidence Level 90% - High Confidence As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps. Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity. |
| 8 | Cash Frenzy - Slots Casino | 60% Confidence Level 60% - Moderate Confidence As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps. Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity. |
| 9 | Vegas Slots - 7Heart Casino | 10% Confidence Level 10% - Low Confidence As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps. Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity. |
| 10 | Slots Casino - Vegas Slot Machine Games | 90% Confidence Level 90% - High Confidence As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps. Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity. |
1
Billion Cash Slots-Casino Game
100%
Confidence Level
100% - High Confidence
As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps.
Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity.
2
Eon Slots Casino Vegas Game
90%
Confidence Level
90% - High Confidence
As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps.
Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity.
3
Cash Tornado Slots - Casino
50%
Confidence Level
50% - Moderate Confidence
As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps.
Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity.
4
Super Vegas Slots Casino Games
44%
Confidence Level
44% - Moderate Confidence
As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps.
Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity.
5
Vegas Nights Slots
40%
Confidence Level
40% - Moderate Confidence
As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps.
Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity.
6
Tycoon Casino - Vegas Slots
90%
Confidence Level
90% - High Confidence
As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps.
Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity.
7
Winning Slots Las Vegas Casino
90%
Confidence Level
90% - High Confidence
As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps.
Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity.
8
Cash Frenzy - Slots Casino
60%
Confidence Level
60% - Moderate Confidence
As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps.
Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity.
9
Vegas Slots - 7Heart Casino
10%
Confidence Level
10% - Low Confidence
As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps.
Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity.
10
Slots Casino - Vegas Slot Machine Games
90%
Confidence Level
90% - High Confidence
As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps.
Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity.
| No | App name | Confidence |
|---|---|---|
| 11 | Epic Vegas Slots - Casino Game | 40% Confidence Level 40% - Moderate Confidence As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps. Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity. |
| 12 | Slots of Vegas | 90% Confidence Level 90% - High Confidence As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps. Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity. |
| 13 | Slots: Vegas Casino Slot Games | 90% Confidence Level 90% - High Confidence As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps. Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity. |
| 14 | Vegas Casino & Slots: Slottist | 88% Confidence Level 88% - High Confidence As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps. Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity. |
| 15 | LuckyBomb Casino Slots | 90% Confidence Level 90% - High Confidence As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps. Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity. |
| 16 | Slots Mega Win Casino Game | 75% Confidence Level 75% - Moderate-High Confidence As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps. Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity. |
| 17 | High 5 Casino: Home of Slots | 90% Confidence Level 90% - High Confidence As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps. Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity. |
| 18 | Winner Slots Casino Games | 20% Confidence Level 20% - Low Confidence As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps. Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity. |
| 19 | Cash Tap Casino: Slot Machines | 90% Confidence Level 90% - High Confidence As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps. Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity. |
| 20 | Cash Tap Casino: Slot Machines | 90% Confidence Level 90% - High Confidence As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps. Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity. |
11
Epic Vegas Slots - Casino Game
40%
Confidence Level
40% - Moderate Confidence
As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps.
Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity.
12
Slots of Vegas
90%
Confidence Level
90% - High Confidence
As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps.
Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity.
13
Slots: Vegas Casino Slot Games
90%
Confidence Level
90% - High Confidence
As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps.
Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity.
14
Vegas Casino & Slots: Slottist
88%
Confidence Level
88% - High Confidence
As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps.
Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity.
15
LuckyBomb Casino Slots
90%
Confidence Level
90% - High Confidence
As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps.
Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity.
16
Slots Mega Win Casino Game
75%
Confidence Level
75% - Moderate-High Confidence
As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps.
Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity.
17
High 5 Casino: Home of Slots
90%
Confidence Level
90% - High Confidence
As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps.
Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity.
18
Winner Slots Casino Games
20%
Confidence Level
20% - Low Confidence
As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps.
Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity.
19
Cash Tap Casino: Slot Machines
90%
Confidence Level
90% - High Confidence
As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps.
Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity.
20
Cash Tap Casino: Slot Machines
90%
Confidence Level
90% - High Confidence
As context is highly nuanced, we also feed our models with the level of confidence in the contextual relationship between both apps.
Confidence is effected by the length of the app description, the quantity of ambiguous words in it, and overall clarity.
Resultados de la agrupación
Ingrese un identificador de agrupación y haga clic en “Calcular” para ver las aplicaciones con el contexto más parecido.
¿Por qué ELMo?
Cuando diseñábamos funciones relacionadas por contexto para nuestros modelos de predicción de instalaciones y compras, tuvimos en cuenta dos ideas clave preconcebidas: 1) el contexto tiene un papel significativo en la predicción de los comportamientos de los usuarios y su interacción con los anuncios; y 2) el contexto puede tener muchos matices.
Estas ideas hacían que usar las categorías de la tienda de aplicaciones no tuviera sentido, ya que, por un lado, introducían un sesgo inherente y, en otros casos, imprecisiones evidentes. Lo anterior se daba a raíz de la optimización de tienda de aplicaciones (ASO) y la implementación de otros enfoques como crear las categorías por nosotros mismos al agrupar mediante métodos de modelado de temas (como LDA o W2V). Como resultado, no generamos la mejora en la obtención de información y precisión que esperábamos. Por estas razones, nos dimos cuenta de que teníamos que encontrar una forma de representar cada agrupación de manera individual.
Después de sopesar la complejidad de la tarea y las herramientas más usadas para solucionar los problemas relacionados con el lenguaje en el aprendizaje automático, el trabajo de investigación de nuestros científicos de datos los llevó a probar ELMo. ELMo usa una red neuronal para aprender las asociaciones entre palabras y sus significados cuando se usan en distintos contextos. Esto lo logra gracias al aprendizaje de las frases en las que dichas palabras se usan dentro de un conjunto de datos gigantesco de 5500 millones de tókenes (palabras y sus partes).
Después de entrenar al modelo ELMo, este representa cada palabra como un vector, lo que nos permite medir la similitud del coseno entre cada palabra. Esto nos indica el grado de similitud semántica y sintáctica entre estas. Debido a que tener en cuenta cada palabra es vital para comprender el contexto, ELMo demuestra ser una de las opciones más confiables para una tarea tan compleja.
Ejemplo de cómo los vectores de palabras están cerca contextualmente mediante la medición eficaz de la similitud del coseno
- Clúster 1
- Clúster 2
- Clúster 3
- Vectores de palabras
- Peso
- Representación de la agrupación
- Modelo de aprendizaje automático
Figura de la formación de un vector de agrupaciones a partir de la suma total de sus palabras
Representación del vector de agrupaciones
Después de generar los vectores de palabras, el siguiente paso es sumarlos en cada agrupación de una forma que represente su contexto real. Presumiblemente, se incrusta toda la descripción de la tienda en una sola representación de vectores, de modo que podamos medir la distancia contextual entre las distintas aplicaciones.
Este paso de nuestro proceso de investigación y desarrollo resultó ser el más complejo. La obtención de promedios a partir de los vectores de todas las palabras en las descripciones de la tienda de aplicaciones suena razonable. Sin embargo, esto implicaría una pérdida valiosa de datos que relacionan la frecuencia con la que aparecen o se repiten las palabras específicas en la descripción particular de la tienda de aplicaciones, así como en todo el corpus.
Después de muchas pruebas, el enfoque que generó la mayor obtención de información durante las pruebas se basaba en un estadístico numérico básico del procesamiento de lenguaje natural o PNL, que se refiere al ámbito del aprendizaje automático que se enfoca en el análisis textual. Dicho estadístico es la frecuencia del término por frecuencia inversa de documento o Tf-idf, que básicamente indica la singularidad de palabras específicas en el corpus.
Al final, nuestra representación de agrupaciones evolucionó a un promedio ponderado de cada vector de palabras multiplicado por su valor Tf-idf normalizado. De esta manera, pudimos enfatizar las palabras menos frecuentes y con más matices que usualmente expresan los temas, mecánicas y funciones singulares de cada aplicación.