152

Data Day Texas - Recommendations

Embed Size (px)

Citation preview

Page 1: Data Day Texas - Recommendations
Page 2: Data Day Texas - Recommendations
Page 3: Data Day Texas - Recommendations

I

Page 4: Data Day Texas - Recommendations
Page 5: Data Day Texas - Recommendations
Page 6: Data Day Texas - Recommendations
Page 7: Data Day Texas - Recommendations
Page 8: Data Day Texas - Recommendations
Page 9: Data Day Texas - Recommendations
Page 10: Data Day Texas - Recommendations
Page 11: Data Day Texas - Recommendations
Page 12: Data Day Texas - Recommendations
Page 13: Data Day Texas - Recommendations
Page 14: Data Day Texas - Recommendations
Page 15: Data Day Texas - Recommendations
Page 16: Data Day Texas - Recommendations

Job recommendations are significantly different

Rapid inventory growth - Millions of new jobs discovered every day

Page 17: Data Day Texas - Recommendations

Job recommendations are significantly different

Rapid inventory growth - Millions of new jobs discovered every day

~ 1.5 million new users visit indeed every day

Page 18: Data Day Texas - Recommendations

Job recommendations are significantly different

Rapid inventory growth - Millions of new jobs discovered every day

~ 1.5 million new users visit indeed every day

Average lifespan of a job is ~30 days

Page 19: Data Day Texas - Recommendations

Job recommendations are significantly different

Rapid inventory growth - Millions of new jobs discovered every day

~ 1.5 million new users visit indeed every day

Average lifespan of a job is ~30 days

One job posting usually meant to hire one individual

Page 20: Data Day Texas - Recommendations
Page 21: Data Day Texas - Recommendations
Page 22: Data Day Texas - Recommendations
Page 23: Data Day Texas - Recommendations
Page 24: Data Day Texas - Recommendations
Page 25: Data Day Texas - Recommendations
Page 26: Data Day Texas - Recommendations
Page 27: Data Day Texas - Recommendations
Page 28: Data Day Texas - Recommendations
Page 29: Data Day Texas - Recommendations
Page 30: Data Day Texas - Recommendations
Page 31: Data Day Texas - Recommendations
Page 32: Data Day Texas - Recommendations
Page 33: Data Day Texas - Recommendations
Page 34: Data Day Texas - Recommendations
Page 35: Data Day Texas - Recommendations
Page 36: Data Day Texas - Recommendations
Page 37: Data Day Texas - Recommendations

Compute similarity

For ui In {Users}

For uj In {Users}

SIMi,j

= compute_similarity(ui,

uj)

Page 38: Data Day Texas - Recommendations
Page 39: Data Day Texas - Recommendations
Page 40: Data Day Texas - Recommendations
Page 41: Data Day Texas - Recommendations
Page 42: Data Day Texas - Recommendations
Page 43: Data Day Texas - Recommendations
Page 44: Data Day Texas - Recommendations
Page 45: Data Day Texas - Recommendations
Page 46: Data Day Texas - Recommendations
Page 47: Data Day Texas - Recommendations

Page 48: Data Day Texas - Recommendations
Page 49: Data Day Texas - Recommendations
Page 50: Data Day Texas - Recommendations
Page 51: Data Day Texas - Recommendations
Page 52: Data Day Texas - Recommendations
Page 53: Data Day Texas - Recommendations
Page 54: Data Day Texas - Recommendations
Page 55: Data Day Texas - Recommendations
Page 56: Data Day Texas - Recommendations
Page 57: Data Day Texas - Recommendations
Page 58: Data Day Texas - Recommendations
Page 59: Data Day Texas - Recommendations

Page 61: Data Day Texas - Recommendations

∩ ∪

Page 62: Data Day Texas - Recommendations
Page 63: Data Day Texas - Recommendations

Items[Ui] = {x

1, x

2, ..x

n}

H

minhashH(U

i)= min{ x∈Items

i| H(x) }

Page 64: Data Day Texas - Recommendations
Page 65: Data Day Texas - Recommendations

Similarity(U1, U2) = 1, if minhash(U1) == minhash(U2)

Similarity(U1, U2) = 0, otherwise

This is an unbiased estimator

Page 66: Data Day Texas - Recommendations
Page 67: Data Day Texas - Recommendations
Page 68: Data Day Texas - Recommendations
Page 69: Data Day Texas - Recommendations

Similarity(U1, U2) = 1, if minhash(U1) == minhash(U2)

Similarity(U1, U2) = 0,

Page 70: Data Day Texas - Recommendations
Page 71: Data Day Texas - Recommendations

Hk

Hk

Prob(minhashH(U

i) == minhash

H(U

j)) = J(U

i, U

j)

Page 73: Data Day Texas - Recommendations
Page 74: Data Day Texas - Recommendations

user → {job1, job2, job3,..}

Page 75: Data Day Texas - Recommendations

H = {H1, H

2, ..H

20}

for user in Users

for hash in H

minhash[hash] = min{x∈Itemsi| hash(x)}

Page 76: Data Day Texas - Recommendations

For ui In {Users}

For uj In {Users}

SIMi,j

= compute_similarity(ui,

uj)

Page 77: Data Day Texas - Recommendations

user1 → (111, 123, 134, 148, ..129)

user2 → (101, 123, 139, 148, ..135)

user3 → (191, 103, 126, 108, ..119)

user4 → (191, 103, 126, 108, ..129)

...

Page 78: Data Day Texas - Recommendations

user → {cluster}

cluster → {users}

Page 79: Data Day Texas - Recommendations

123 → (user1, user2)

148 → (user1, user2)

129 → (user1, user4)

191 → (user3, user4)

...

Page 80: Data Day Texas - Recommendations

Page 81: Data Day Texas - Recommendations

user1 → {job1, job2}

user2 → {job2, job3, job5}

123 → {user1, user2}

Page 82: Data Day Texas - Recommendations

user1 → {job1, job2}

user2 → {job2, job3, job5}

123 → {job1, job2, job3, job5}

Page 83: Data Day Texas - Recommendations

1. user → {cluster}

Page 84: Data Day Texas - Recommendations

user → {cluster} user1 → {111, 123, ..}

Page 85: Data Day Texas - Recommendations

111 → {job5, job2, job9}

123 → {job1, job2, job3, job5}

{job2, job5, job9, job1, job3}

→→

Page 86: Data Day Texas - Recommendations

{job2, job5, job9, job1, job3}

Page 87: Data Day Texas - Recommendations
Page 88: Data Day Texas - Recommendations
Page 89: Data Day Texas - Recommendations
Page 90: Data Day Texas - Recommendations
Page 91: Data Day Texas - Recommendations

1.

Page 92: Data Day Texas - Recommendations

→ {101, 121}

Page 93: Data Day Texas - Recommendations

→ {101, 121}

{“Software Engineer”,

“Java Developer”, “Python Developer”}

Page 94: Data Day Texas - Recommendations

→ {101, 121}

{“Software Engineer”,

“Java Developer”, “Python Developer”}

minhash({“Software Engineer”, “Java Developer”,

“Python Developer”}) → {99, 135}

Page 95: Data Day Texas - Recommendations

→ {101, 121}

{“Software Engineer”,

“Java Developer”, “Python Developer”}

minhash({“Software Engineer”, “Java Developer”,

“Python Developer”}) → {99, 135}

→ {99, 121}

Page 96: Data Day Texas - Recommendations
Page 97: Data Day Texas - Recommendations

minhash({“Software Engineer”, “Java Developer”,

“Python Developer”}) → {99,121}

99 → add {“Software Engineer”, “Java Developer”,“Python Developer”}

121 → add {“Software Engineer”, “Java Developer”,“Python Developer”}

Page 98: Data Day Texas - Recommendations

{“Software Engineer”, “Java Developer”,

“Python Developer”} {99, 121}

99 → {“Software Engineer”, “Java Developer”, “Python Developer”}

Page 99: Data Day Texas - Recommendations

→ {99, 131}

{“Software Engineer”, “Java Developer”,

“Python Developer”}

Page 100: Data Day Texas - Recommendations
Page 101: Data Day Texas - Recommendations
Page 102: Data Day Texas - Recommendations
Page 103: Data Day Texas - Recommendations
Page 104: Data Day Texas - Recommendations
Page 105: Data Day Texas - Recommendations
Page 106: Data Day Texas - Recommendations
Page 107: Data Day Texas - Recommendations
Page 108: Data Day Texas - Recommendations
Page 109: Data Day Texas - Recommendations
Page 110: Data Day Texas - Recommendations
Page 111: Data Day Texas - Recommendations
Page 112: Data Day Texas - Recommendations
Page 113: Data Day Texas - Recommendations
Page 114: Data Day Texas - Recommendations

Page 115: Data Day Texas - Recommendations
Page 116: Data Day Texas - Recommendations

●●

Page 117: Data Day Texas - Recommendations
Page 118: Data Day Texas - Recommendations
Page 119: Data Day Texas - Recommendations

1. http://go.indeed.com/docservice

Page 120: Data Day Texas - Recommendations
Page 121: Data Day Texas - Recommendations
Page 122: Data Day Texas - Recommendations
Page 123: Data Day Texas - Recommendations

Page 124: Data Day Texas - Recommendations
Page 125: Data Day Texas - Recommendations
Page 126: Data Day Texas - Recommendations
Page 127: Data Day Texas - Recommendations
Page 128: Data Day Texas - Recommendations
Page 129: Data Day Texas - Recommendations
Page 130: Data Day Texas - Recommendations
Page 131: Data Day Texas - Recommendations

Page 132: Data Day Texas - Recommendations
Page 133: Data Day Texas - Recommendations
Page 134: Data Day Texas - Recommendations

Page 135: Data Day Texas - Recommendations
Page 136: Data Day Texas - Recommendations
Page 137: Data Day Texas - Recommendations
Page 138: Data Day Texas - Recommendations
Page 139: Data Day Texas - Recommendations
Page 140: Data Day Texas - Recommendations
Page 141: Data Day Texas - Recommendations
Page 142: Data Day Texas - Recommendations
Page 143: Data Day Texas - Recommendations
Page 144: Data Day Texas - Recommendations
Page 145: Data Day Texas - Recommendations
Page 146: Data Day Texas - Recommendations
Page 147: Data Day Texas - Recommendations
Page 148: Data Day Texas - Recommendations
Page 149: Data Day Texas - Recommendations

●●●

Page 150: Data Day Texas - Recommendations

Engineering blog & talks http://indeed.tech

Open Source http://opensource.indeedeng.io

Careers http://indeed.jobs

Twitter @IndeedEng

Page 151: Data Day Texas - Recommendations
Page 152: Data Day Texas - Recommendations