fix: major classification uses word-level matching, not substring

'stat' no longer matches 'state', 'cs' no longer matches 'statistics'.
Both _classify_major (contribute) and _profile_adjustment (prediction)
now match against individual majors using word splitting.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: YichengYang-Ethan

cli/main.py

@@ -330,14 +330,22 @@ def _classify_major(majors: list) -> str:
     """Map majors to anonymous category."""
     if not majors:
         return "Unknown"
-    combined = " " + " ".join(majors).lower() + " "
-    quant_kw = ["math", "stat", "physics", "computer"]
-    n_quant = sum(1 for kw in quant_kw if kw in combined)
+    _QUANT = [
+        "math", "mathematics", "applied math",
+        "statistics", "stats", "stat",
+        "physics",
+        "computer science", "computer", "computing", "cs",
+    ]
+    _ECON = ["econ", "economics", "finance", "financial"]
+    n_quant = sum(
+        1 for m in majors
+        if any(kw == m.lower() or kw in m.lower().split() for kw in _QUANT)
+    )
     if n_quant >= 2:
         return "Multi-quant (math/stats/CS)"
     if n_quant >= 1:
         return "Quant-related"
-    if any(kw in combined for kw in ["econ", "finance"]):
+    if any(any(kw in m.lower() for kw in _ECON) for m in majors):
         return "Econ/Finance"
     return "Other"
 

core/lr_predictor.py

@@ -231,14 +231,25 @@ def _profile_adjustment(profile: "UserProfile") -> float:
     # --- Major relevance (+0 to +0.15) ---
     majors = getattr(profile, "majors", [])
     if majors:
-        major_lower = " ".join(majors).lower()
-        quant_kw = ["math", "stat", "physics", "computer", "cs"]
-        n_quant = sum(1 for kw in quant_kw if kw in major_lower)
+        _QUANT_MAJORS = [
+            "math", "mathematics", "applied math",
+            "statistics", "stats", "stat",
+            "physics",
+            "computer science", "computer", "computing", "cs",
+        ]
+        _ECON_MAJORS = ["econ", "economics", "finance", "financial"]
+        n_quant = sum(
+            1 for m in majors
+            if any(kw == m.lower() or kw in m.lower().split() for kw in _QUANT_MAJORS)
+        )
         if n_quant >= 2:
             adj += 0.15
         elif n_quant >= 1:
             adj += 0.08
-        elif any(kw in major_lower for kw in ["econ", "finance"]):
+        elif any(
+            any(kw in m.lower() for kw in _ECON_MAJORS)
+            for m in majors
+        ):
             adj += 0.05
 
     return adj